Relative Intensity & Training To Failure

When we talk about intensity, we are generally talking about how “heavy” a weight is – i.e. what % of your one repetition max it is. So, a heavy set of 3 would be considered to be “high intensity”, as it involves you using a weight that is very close to the most amount of weight you could possibly lift. However, there is another “intensity” that concerns us – relative intensity. Relative intensity can be thought of as your intensity of effort. What this means is that a high relative intensity would reflect the performance of close to the most amount of repetitions you could do with a given weight. On the other hand, your warm-up sets would be considered to be at a low relative intensity, since you are not doing anywhere near as much as you possibly could at that load.

To the novice trainee, relative intensity may seem like a worthless concept. Fitness culture would often lead one to believe that there is no option but to give 100% all of the time. However, note that I said, “to the novice trainee”. Although I am unsure where the idea originated, one of the soundbites that circulates in the exercise science world is the idea that the sub-elite athlete will have mastered putting in all of his/her effort, whereas the elite athlete will have mastered the appropriate allocation and management of effort. What this alludes to is the fact that giving 100% is not always desirable. Regardless of whether your goal is general health, hypertrophy, strength or endurance, one has to recognise the difference between building fitness and expressing fitness. If you want to develop, you need to focus on investing, not cashing in. 

Relative Intensity Origins & Current Concepts

Clearly, it is not enough to have a trainee say that a set was “hard”. In order to have some sort of quantifiable measure to be used in clinical, research and sporting settings, there has to be some sort of scale to quantify exactly how easy or hard a given exercise bout was. One such method of doing so is the “Rate Of Perceived Exertion” (RPE) scale. Contrary to what you may be aware of, the version you may see used these days in the weight training setting is actually a modified version. The original RPE scale was devised and tested by Gunnar Borg, which was a 15-point scale, where the subject would give a rating from 6-20, reflecting how challenging an exercise bout was. The intention of using 6-20 was to translate these readings into heart rate estimates, where a 6 would reflect a resting heart rate (60bpm), whereas a 20 would reflect a max heart rate (200bpm). A 6 was used to describe no exertion at all i.e. at rest, a 9 would be very light exertion, 11 light, 13 somewhat hard and so on up to 20, which would be maximal exertion. He conducted an abundance of research in this area from the 1960s – 1990s and deserves to be credited for his work in this area, as he is sometimes forgotten about in contemporary discussions of the modified scales. 

After the conception of the initial 15 point scale, he also created the Borg Category Ratio Scale (CR-10), which was a 10-point version used to quantify dyspnoea (shortness of breath) in the clinical setting. This is still used today in multiple populations, including respiratory disease, cardiovascular disease etc. So, he did actually create a 10-point scale too, not just the 6-20 scale.

Furthermore, powerlifter and coach Mike Tuchscherer of Reactive Training Systems is suggested to be the first to implement the RPE scale for resistance training. He adopted the 0-10 scale, although I have read elsewhere that the 0-10 scale has only been used in research papers since Mike’s conception of it, which is false, as the CR-10 scale developed by Borg had already been used prior to this. Borg’s RPE scale had also been used in research prior to Mike’s edition, which is often misstated on blogs on the internet, which suggest that rate of perceived exertion had not been used in relation to resistance training prior to this. Likewise, it seems some people also use Mike’s version of the scale, while crediting it to Borg, so basically, we need to get our story straight and give credit where credit is due lol. 

Anyway, Mike’s use of the RPE scale was based not on breathlessness or arbitrary descriptors of “easy” or “hard”, but rather on how many repetitions in reserve someone felt they had at the end of a set. The modified RPE scale, which estimates RPE based on the number of repetitions in reserve, was initially played out by Mike in the Reactive Training Systems Manual (Tuchscherer 2008) and has been heavily researched in the resistance training world since, most notably by Dr Mike Zourdos and colleagues. 

So, essentially, when we talk about either RPE or RIR, we are referring to the above scale, where the RPE rating corresponds to a corresponding prediction of the number of RIR. Regardless of whether you choose to measure and note RPE or RIR in your training, they are essentially measuring the same thing.

Classically, resistance training has been prescribed based on percentage (%) guidelines, typically based on the % of one’s one repetition maximum (1RM). For example, a programme may read “5 sets of 5 repetitions at 75% 1RM”. Within this framework, the intent to manage relative intensity can still be present, as the % guideline given generally comes with an assumed number of reps one could perform with that load. For example, in the case of the chart below, if one was to prescribe 5 x 5 at 75% 1RM on a programme, they would be doing so under the assumption that the trainee could probably achieve up to 10 repetitions with that weight. Therefore, they will likely have between 3-5 RIR on all of those sets, or, be working at an RPE of 5-7, depending on the length of their rest periods and how much fatigue they carry from set to set. 

Therefore, accounting for relative intensity is not synonymous with using RIR or RPE guidelines, as it is still indirectly factored into % based guidelines. The RPE exists whether you estimate/measure it or not, which is one of the reasons it may be a good idea to do so (at least for a period of time).

Why Not Just Use Percentages Rather Than Relative Intensity?

The big difference here is that the intent of accounting for relative intensity through the use of % guidelines fails to acknowledge the fact that one’s 1RM is not a static number on a day-to-day basis. This can work both ways:

• If someone tested their 1RM at a point in time where they were very fatigued, low on sleep, highly stressed and in a caloric deficit, then subsequent training recommendations based on that 1RM would underestimate the trainee’s ability and hence the relative intensity would be lower than predicted.
• On the other hand, if someone tested their 1RM on a day where their mood and motivation was high, they were well fed, had slept well and were carrying a low amount of fatigue, then subsequent training recommendations may overestimate the trainee’s ability and hence the relative intensity would be lower than predicted.

From this, we can conclude that in order to apply a specific training stress, we need to account for other variables and hence we need to recognise that 1RM is not a static point. If 1RM is not a static point, then the training stress imposed by a given % of that number is going to vary from session to session. Therefore, using one’s perception of that stress on that day ± the eye of a coach or video feedback, RPE / RIR can be noted and used to select training loads.

If you are provided with a 12-week programme that has a built in progression scheme based on percentages, it may look something like this:

Week 1 – 3 x 6 @ 80%
Week 2 – 3 x 6 @ 82.5%
Week 3 – 3 x 6 @ 85%
Week 4 – 3 x 6 @ 87.5%

The problem with this approach is that it falls victim to epistemic arrogance*, where one assumes to know far more than they actually do. It is impossible for any human to know at what rate a trainee is going to progress, nor could they be prospectively aware of the status of all of the variables that could impact performance on a given day.

Therefore, by prescribing load/progression based on percentages in isolation**, one runs the risk of grossly over/under-dosing the training stimulus on a given day. This could mean working at a low level of effort on the days where you are most prepared, but at a high level of effort on the days where you are least prepared. By using RPE to guide load selection, these variations can be dealt with autonomously by adjusting loads based on perceived effort on a given day.

*Epistemic arrogance: ”Measure the difference between what someone actually knows and how much he thinks he knows. An excess will imply arrogance, a deficit humility. An epistemocrat is someone of epistemic humility, who holds his own knowledge in greatest suspicion” (Taleb 2007).

**One can also combine RPE & percentage-based prescription, the application of which will be discussed later on.

Failure vs Non-Failure For Hypertrophy & Strength

Before we go any further into the weeds of using RPE/RIR, we need to address the elephant in the room. This elephant is, of course, the beast mode attitude that prevails in the fitness world, which would suggest that no less than 100% effort is acceptable. Along with that, there is the other extreme camp, who feel that going to failure is always detrimental and probably stay far too shy of that point. As always, the middle ground is probably where the best answer is, so, to get there, let’s dig into some of the details of training to failure.

The Motor Unit Argument

One of the primary arguments for training to failure is that it activates all motor units* and hence muscle fibres. This brings us back to the fundamental understanding of what drives adaptations. With mechanical tension recognised as the primary driver of hypertrophy, it would seem intuitive that heavier loads = more tension = more hypertrophy. And to a degree, that is true, as once a load is heavy enough, it will lead to the recruitment of all motor units. More specifically, once we reach around 80% of our max (specifically, our maximum voluntary isometric contraction (MVIC)), all motor units will be active from the beginning of the set (Pinto et al. 2013). There is of course going to be inter-individual variation in the intensity required for all motor units to be recruited, as is the case with a lot of resistance training concepts, but for the purpose of this article, that is relatively irrelevant. From this, we understand that once we are using a heavy load, we are already recruiting the full spectrum of motor units.

*A motor unit is a group of muscle fibres innervated by the same motor nerve. They vary in their recruitment threshold, meaning that some motor units require higher force demands to be recruited. Because motor unit recruitment is demand-driven, the nervous system will only recruit as many motor units as are required to produce the force required. The smaller, low-threshold motor units will be recruited first, followed later by the high-threshold motor units as required. This is known as Henneman’s Size Principle. As we will see, this is also impacted by fatigue.

Some of you who may be a little more savvy with muscle physiology may be aware that we can maximise motor unit recruitment by simply lifting explosively, and while this is true, it does not tell the whole story. Mechanical tension is not synonymous with motor unit recruitment, as there needs to also be sufficient “cross-bridges” (essentially the connections between the actin and myosin proteins**) within the muscle fibre to maximise mechanical tension within individual fibres and hence, the hypertrophic response. 

**interestingly, we are currently seeing a low key revolution ongoing regarding our understanding/lack of understanding of muscle contraction, so we may see theories that overrule cross-bridge theory, but anyway, irrelevant for now.

Now, we understand heavy loads to be effective via their ability to maximise motor unit recruitment and be sufficiently challenging to reduce concentric velocity and hence maximise the tension in the actin-myosin cross-bridges. However, what this fails to explain is the relationship between failure vs non-failure training and light loads. You would be right in thinking that loads as light as 30% 1RM are simply not heavy enough to demand the full spectrum of motor units to be recruited [in the absence of fatigue]. In this case, metabolic stress and fatigue play a role. Due to the nature of light weight sets requiring very high repetitions, there is an increased accumulation of metabolites within the muscle. This metabolic stress contributes to the fatigue of the active fibres. As the involved fibres begin to fatigue, those fibres are no longer sufficient to produce the mechanical tension required to move the weight, meaning that higher threshold motor units are recruited to help out. This occurs in accordance with Henneman’s Size Principle, which means that the slow twitch fibres, which produce less force but are more fatigue resistant, will be recruited first, followed later by the fast twitch fibres, which produce higher forces but fatigue quicker. What this means is that as the set progresses and you approach failure, you recruit the full spectrum of motor units, just like you would with a heavy load. And because the fatigue leads to a reduction in concentric velocity and hence more actin-myosin cross-bridges, that also means we have both conditions required to maximise the mechanical tension on individual muscle fibres. Basically, with heavier loads, you’re already recruiting all of your motor units and the velocity of your concentric will be slow, however, for lighter loads, when you approach failure, the velocity will then begin to reduce and the remaining motor units will begin to be recruited. 

Therefore, motor unit recruitment and proximity to failure is an intensity-dependent relationship. Put very simply, the harder a given repetition is, the more likely it is that you are at a point where you are either recruiting most, if not all, of your motor units OR, you have already used up the lower threshold motor units, incurred fatigue and are now at a point where the higher threshold motor units are doing their job. With that said, we can conclude that the emphasis on pushing close to failure is more important for lighter loads, since we need to be approaching failure to achieve a similar training effect. However, this doesn’t mean anything above 5 reps – if we are talking about motor unit recruitment, then we have evidence to suggest that when training at ~15RM, muscle excitation will be maximised 3-5 reps from failure, therefore you are likely in a good place once you start to get within that range (Sundstrup et al. 2012). Within 3-5 reps of failure in that rep range, the concentric velocity is also likely to be reducing, increasing per-rep mechanical tension on the active fibres, so I think the light weight all the way to failure “need” is probably more relevant for sets of 20+.

Failure & Fatigue

Having said that, we do also need to think about how the resulting fatigue from training to failure may affect muscle recovery and hence one’s ability to provide another training stimulus in the days to follow. This is part of looking at the bigger picture. If we choose to look solely at a single repetition’s training effect, a repetition that is more challenging is, of course, going to provide a greater training stimulus. However, with a more potent stimulus, we get more fatigue, which can affect the total amount of work we can do that day, week, month etc. Thankfully, we do have some research investigating such variables.

In this study, the authors aimed to describe the recovery response to exercise, as it relates to the proximity of repetitions to the maximum achievable. Of interest to us, this essentially describes recovery when training to failure (RPE 10 / RIR 0) vs leaving reps in the tank. 

10 highly resistance-trained males (average resistance training experience of 8.2 ± 3.5 years – however the average bench press 1RM was 87.2 ± 15.2kg, which doesn’t sound like 8 years of training, but anyway) took part in the study, completing three different protocols, all of which consisted of a bench press and squat performed in the smith machine:

Low volume, low relative intensity – 3 sets of 5 reps with 75% 1RM.

High volume, low relative intensity – 6 sets of 5 reps with 75% 1RM.

High volume, high relative intensity – 3 sets of 10 reps with 75% 1RM.

Conditions 2 and 3 are really what are of interest to us for comparison, as they are “volume-equated”, meaning that the total volume load (sets x reps x load) is equal between the two conditions, with the only variable being the relative intensity. The rest-periods were also standardised at 5 minutes between sets. Each condition was performed by each subject with 4 weeks between tests, which would likely minimise any repeated-bout effects that could be protective against subsequent training stressors. The study also had some other cool features, including standardised nutrition, accommodation in the facility and other controlled variables, in the 5 days of each experiment (day of, day before and 3 days after). You can check out the details in the study, as I don’t want to bore you further.

The following measures were used as markers of recovery:

Neuromuscular Fatigue – Countermovement jump height (light load), performance with a load that initially could be moved at 1m/s (medium load) and performance with 75% 1RM (high loads). These were intially assessed pre-exercise and again assessed directly post, 6 hours post, 24 hours post, 48 hours post and 72 hours post.

Biochemical Measurements – Serum total testosterone, cortisol, growth hormone, creatine kinase and ammonia were all measured at the same time-points via blood samples

To get the boring results out of the way, no significant differences in any of the above variables were noted between conditions 1 and 2, both of which were far from failure, with the exception of growth hormone directly after the training, which was higher in condition 2. To dig into the weeds of every specific measure and time-point, you can take a look at the graphs from the study, which I have attached. However, if that bores you, here are the main results of interest for this discussion i.e. changes in markers of fatigue/recovery in the volume-equated conditions (condition 2 vs 3):

• Condition 3 (3×10) resulted in a significantly greater acute decline in performance (at 0h post, measured by the aforementioned neuromuscular performance measures). 
• The two non-failure conditions (conditions 1 & 2) resulted in faster mechanical recovery at 24 and 48h post, meaning a new training session could be performed.
• The blood markers also reflected this, with the failure condition resulting in higher acute fatigue (as measured by growth hormone and ammonia post-exercise), along with delayed markers (creatine kinase – a marker of muscle damage), highlighting that recovery can take up to 24-48 h.

From this, what we can take away practically is that for a given amount of total work, the resulting fatigue will vary depending on proximity to failure. Clearly, this would push us in the direction of concluding that it’s probably a good idea to stay shy of failure, as you can complete more total work with less fatigue. This is further supported by more recent research by Santos et al. (2019), who found that the increase in repetitions on earlier sets in the failure group was compensated for by the reduction in reps in later sets;

“Compared with non-failure, momentary failure resulted in a higher number of repetitions in the first set (11.58 6 1.83 vs. 7.58 6 1.72, p , 0.05), but a lower in the last set (3.58 6 1.08 vs. 5.41 6 1.08, p < 0.05). Total number of repetitions was similar between the protocols (MF 26.25 6 3.47 vs. NF 24.5 6 3.65, p < 0.05).”

So, while you may get more repetitions initially, the fatigue can limit the performance of subsequent sets, resulting in similar total repetitions performed across a fixed number of sets. This fatigue and effort perception was also measured via 1) RPE, 2) session RPE, and 3) rate of perceived discomfort, all of which were higher in the failure group. Hence, training to failure resulted in similar total reps, but with more fatigue.

However, we have to consider 1) whether or not there are differences in strength outcomes based on proximity to failure, 2) whether or not there are differences in hypertrophy outcomes based on proximity to failure and 3) if there are differences in outcomes, what RPE/RIR we should use as a cut-off that approximates a good trade-off of additional fatigue for those outcomes. As always, this is not about giving you a black-and-white answer, it’s about giving you an appreciation for the nuance that can help you inform your programming going forward.

Strength & Hypertrophy Outcomes

This meta-analysis, which had initially suggested that non-failure training was superior, has recently been updated to suggest that there is simply no statistically significant difference in strength outcomes between failure vs non-failure training. In untrained subjects, the strength increase was ~34% between both groups, suggesting that subtle differences in programming, such as relative intensity changes, are unlikely to meaningfully affect strength outcomes. Furthermore, the authors had initially suggested that non-failure training was superior for trained subjects, since they experienced a ~14% increase in strength vs the ~12% increase seen in those who trained to failure, however, in their erratum to the paper, they suggested that this be deleted, with the general tone of their paper now reflecting the perspective that there probably isn’t a meaningful difference in outcomes. They do also acknowledge the differences in fatigue incurred between failure vs non-failure training, suggesting that non-failure training may allow for greater incremental load increases throughout the course of a training programme. 

However, it’s not really enough to just look at failure vs non-failure, as we do need to take a look at how different proximities to failure affect outcomes in both volume-matched and non-volume-matched situations. When I am talking about it being “volume-matched”, I am referring to an equal volume load, which is sets x reps x load. For example, 10 x 10 x 100kg = 10,000kg volume load. This can give a better insight into the role of failure, independent of differences in total volume load. The reason this is important is because if we just say 3 sets to failure is better than 3 sets shy of failure, then there is clearly a difference in total work done and hence it’s not really a fair comparison. It’s sort of like asking “do bigger meals lead to greater weight gain?”. Obviously, if you compare 3 big meals vs 3 small meals, then yes, but if someone eats 5 small meals instead of 3 big meals, it’s not as clear-cut. 

In this study, the authors examined 3 different conditions over a 10-week period, similar to those in the fatigue study, however, the goal here was to investigate the effects on hypertrophy and strength. The 3 training groups were as follows:

Repetitions to failure (RF) – 3 sets to failure at 70% 1RM.

Repetitions not to failure, but with equal volume (RNFV) – 4 sets of 7 reps at 70% 1RM.

Repetitions not to failure (RNF) – 3 sets of 7 reps at 70% 1RM.

The drawback of the study was that these were untrained, physically active young women, therefore we have to be careful not to conclude that this applies to everyone. It was also a bicep-focused training study, meaning that we may see different outcomes if applied to the rest of the body, although they did simultaneously train the rest of the body, just not for the purpose of outcome measurement. Regardless, we do have to accept that all studies will have limitations. 

From the results, we can see that the RF group did end up performing the same amount of training volume (statistically insignificant difference), which was the goal of the study. However, there was a difference in total volume was in the RNF group, which was expected and desired. 

In terms of the outcomes of interest, there were no significant differences in maximal strength or muscular endurance, with the only significant difference in power being that the RF group experienced impaired force production at higher velocities when peak torque was assessed at 180 degrees per second – this may have implications for those looking to express their strength at high velocities in a sports context. In terms of hypertrophy, the results suggest that muscle thickness increased by 17.5% in the RF group vs 8.5% in the RNFV group, however, the authors suggested that there was no difference, eluding to it being non-statistically significant. In essence, this study highlighted that the biggest predictor of both strength and hypertrophy outcomes over a 10 week period in these specific subjects was not whether or not sets were taken to failure, but rather the total volume performed. Our conclusions are also limited by the fact that there is not a clear explanation as to how load increases were made, therefore, the RF group could have been increasing the loads they used sooner and were always training to do as much as they could, whereas the RNFV group could have had progressively easier sessions until the midpoint of the study, when loads were increased after 1RM testing.

So, what I would take away from this study is that the RF group had the greatest increase in muscle mass, whereas strength increases showed no difference. We should also keep in mind that the RIR in the non-failure group was likely no more than 2-3 RIR. The reason we can take this away is because the volume was supposedly matched between 28 reps (4×7) and 3 x failure, meaning that the failure group likely got between 9-10 reps. Therefore, this failure vs non-failure condition is more relevant than those studies that compare sets to failure vs those with 5+ RIR, as this is more likely to be what people do in the real world. 

Furthermore, we do have more research than just this, most of which seems to point in the opposite direction regarding hypertrophy. While I don’t want to break down every study in detail, consider the following paper by Sampson & Groeller (2016), who carried out a 12-week study in 28 males, designed to investigate the effects of failure (with a 2s concentric and eccentric) versus non-failure in two different conditions (rapid concentric, controlled eccentric OR rapid concentric, rapid eccentric) on strength, hypertrophy and muscle excitation. “A significant increase in 1RM (30.5%), MVC (13.3%), CSA (11.4%), and agonist EMG(RMS) (22.1%) was observed; however, no between-group differences were detected.” Therefore, similar adaptations took place in each condition, even though the failure group performed significantly more repetitions – implying more work and more fatigue for the same adaptations.

Essentially, we have a situation whereby our theory and the evidence both come together to suggest that pushing your sets all the way to failure probably isn’t necessary for strength or hypertrophy. Not only that, but in the context of a full resistance training programme, the additional fatigue accumulated by pushing all the way to failure may reduce your hypertrophy/strength outcomes by reducing total volume, having a more prolonged reduction in force output and reducing the potential frequency at which you could train, along with increasing muscle damage. Therefore, it’s probably a robust recommendation to suggest that training to failure should be reserved for higher repetition sets with lighter weights, as opposed to higher intensity work. And even if you are working with lighter weights, you have to consider how going to failure may affect the total work you have planned that day/week. It’s likely to be safe to train to failure, and it’s also unlikely you will “overtrain” or meaningfully compromise your results by doing as many reps as you can, but you just have to look at the bigger picture and recognise that conservation of effort may be beneficial to manage fatigue and potentiate your ability to do more work overall over a given unit of time (e.g. 1 week, 1 month, 12 weeks etc.).

To Play Devil’s Advocate…

One of the points that needs to be made in favour of pushing to failure, at least some of the time, is that it can teach you where your repetition maximum with a given load actually is. If you never work close to failure, then it is unlikely you will be able to gauge how many reps you have left in the tank and hence you could misinterpret your relative intensity. This is especially important given the variety of responses we see when people push toward failure – some people go red and look like they are going to explode, some people look like they are meditating, some continue to move loads quickly, but some can grind out at a snail’s pace. With that in mind, relying on how something looks when you watch it back on video may not be all that useful, unless you actually know what it looks and feels like to work close to failure in the first place.

You may be thinking “come on, everyone knows how much they can lift”. You’d be wrong…

This study looked at what people reported to be the load they normally choose to perform for 10 reps on the bench press, then got them to perform as many reps as they could with that weight, with the average reps achieved actually being 16 repetitions (±5 reps, meaning the range was 11-21). Therefore, if these subjects were to go on a programme that suggests they stay 2 reps from failure, they may have skewed perceptions as to what they can really do. However, this is, of course, subject to inter-individual variability. 22% of the subjects in that study could in fact only get 10-12 repetitions, meaning that ~ 1 in 5 trainees likely trained within a couple of repetitions of their maximum repetitions regularly. 31% were able to do 13 to 15, 21% could do 16-18 and a whopping 26% of subjects could perform 19-20+ (~14% greater than 20). Interestingly, 76% of the subjects reported hypertrophy as a primary goal of their training, and there was no effect of the subjects’ goals on the outcome in terms of how many reps they got and hence, how effective they were at self-selecting appropriate loads.

With that in mind, we can assume that there is a large proportion of recreational lifters in your gym who really don’t know where their strength lies. If you never test, you may not know. Therefore, I think an argument can probably be made for seeing where failure is every now and then, as it will give you a feel for what a set actually feels like when you are doing as much as you possibly can. This is especially important if you find that the weight you normally do for 8-10 reps is actually a 15-rep max for you when you really push yourself. You probably don’t want to be ending your sets with 5-7 reps in the tank if you are trying to build muscle.

Furthermore, the good news is that one’s ability to predict performance increases with resistance training experience, as shown by Steele et al. (2017) in their paper investigating trainees’ competence in predicting repetitions to momentary muscular failure. The subjects were categorised as follows based on training experience: orientation (<1.5 months), beginner (1.5-6 months), experienced (6-12 months), advanced (12-36 months) and expert (>36 months). The subjects then carried out a resistance training programme, inclusive of exercises that they had to have previously been performing as part of their training. They were asked to predict how many reps they could perform until failure and then followed it up by testing that. In short, the more experienced the lifter, the better their predictions were. Here is an isolated example extracted from the results: 

On the chest press orientation level lifters predicted 15.40 ± 2.77 reps, but performed 20.47 ± 6.36 reps, which would be a 5 RIR if they normally perceived 15.40 ± 2.77 to be their max. However, expert level lifters predicted 10.48 ± 2.86 reps and performed 10.93 ± 3.17 reps, which is bang on. This trend was consistent through all categories, with a 4.93 rep discrepancy at orientation level, 3.9 at beginner, 1.86 at experienced, 1.03 at advanced and 0.45 at expert.

With these pieces of evidence considered, it’s important to consider the individual before recommending relative intensity guidelines. If someone is told to leave 3 RIR, when they already fall shy of what they can do anyway, then they are going to be pretty far from failure and really missing out on the hypertrophy benefits of approaching failure. Therefore, if you are a coach, then showing clients where they can take their sets to may be a means of speeding up this learning process. However, it’s likely to come with time anyway, so it may just be a case of encouraging the performance of as many reps as possible every now and then to ensure they really are pushing themselves in training, as opposed to instilling a mindset of being over-conservative.

Basically, this section is essentially me scientifically saying “people just don’t train hard”. The thing that is ironic about all of this is that people often think that basing training decisions on scientific evidence is all about making training easier and not doing the work, whereas if you actually dig into the research, the clear implication is that most people probably don’t train hard enough or when they do, they simply misplace their efforts. At the end of the day, humans are not perfect and we tend to want to make things easier for ourselves when we get uncomfortable, so it’s no different with training. To reinforce this idea and really finalise our point here, consider how RPE ratings vary with lower-body training, which is generally a lot less comfortable than upper-body training. When we rate RPE or RIR, we are referring to perceived effort vs the max that we could do, NOT a sensation of discomfort. This is one of the major criticisms of rating RPE/RIR, as the sensations associated with exercise can confound the reliability of effort/exertion ratings.

One further criticism supported by the literature is the idea that RPE/RIR is less accurate for higher rep sets. Anecdotally, this has a lot of support. I’m sure you have experienced a drastic jump in reps attained when you really push to failure with a weight you were previously using for sets of 10-12. This was investigated by Zourdos et al. (2019), who found that RIR predictions were less accurate when subjects performed more reps (e.g. 10 reps @ 70% 1RM vs 15 reps @ 70% 1RM), and for ratings further from failure (e.g. 5 RIR vs 1 RIR). Essentially, the further you are from failure, and the lighter the load/higher the reps, the more difficult it is to accurately gauge the amount of RIR.

Application & Practical Strategies

With all that said, now it’s time to explore how you can actually use RPE/RIR to guide your training, as it’s all very well to get stuck into the theory, but at the end of the day, it’s all there to influence the decisions you make when writing training programmes.

Relative Intensity & Load Prescription – with Repetition Target

You could use a load range (e.g. 100kg ± 5kg) and then specify your load selection by gauging it against the RIR target. This is essentially the foundation of using RIR: you use it to guide the loads that you used. Most programmes that use RIR will use it in this way. For example: 4 x 3 @ 2 RIR. That’s generally how you will see it applied, as this allows you to specify the intensity (sets of 3 at your 5RM, more likely to drive strength adaptations), while still ensuring fatigue is being managed (not going to failure). 

Relative Intensity & Load Prescription – without Repetition Target

Although it may not be what you are used to, you could write a training programme without a specific repetition target. Instead, you could use a relative intensity guideline to give a defined point at which the set ends. To put this into practice, you could suggest performing a certain load (e.g. 100kg) at 2 RIR. You could then simply approach each of your sets with the intent of working to the point where you feel you have no more than 2 RIR. This may be the same number of reps on each set, or it could vary. That’s the beauty of RIR – it allows you to have some flexibility while still trying to specify the training stress applied.

Another way you could do this would be to have a broad range (e.g. 12-20) and an RIR cut-off, without any recommended load selection. This is especially useful for more isolation-type exercises in higher repetition ranges, as people generally don’t have the same mental tie to specific loads on those exercises and are more open to adjusting their loads based on the relative intensity. For example, you could programme 12-20 reps at 1-2 RIR, which can then drive subsequent load selection.

Relative Intensity & Programme Progression – with Increasing Relative Intensity

If you use the reps in reserve (RIR) method of quantifying your relative intensity, then you could progress that relative intensity as you move through a certain phase of programming. For example, you could have an 8-week block of training where you increased your effort weekly, so that RIR decreased from 3 on all sets on week 1 to 1 on all sets on week 8. This could then guide either the load increases you make each week OR the rep target increases that you make. The good thing about using RIR to guide your progression is that it is more specific to your performance on the day than pre-determined load increases. For example, if on week 6, you are due to work with 2 RIR on all sets, but you actually feel a little fatigued on that day, so that lighter weights than normal are leaving you with 2 RIR, then you can simply accept that for what it is and recognise that that is the level of training stress that will align with the 2 RIR training stress you had planned. The load may be different and may not have progressed, but your effort is still where it is supposed to be, as opposed to going ahead with increasing the load anyway and further increasing the amount of fatigue you accumulate that week.

Relative Intensity & Programme Progression – with Fixed Relative Intensity

In this case, you can use the RIR guidelines to guide the way in which you choose loads or repetitions, much like you would if you were increasing relative intensity. In a fixed relative intensity model, you may look to have a consistent level of effort each week, which is often something that is useful for beginners, who can progress pretty quickly and hence a small increase in load may be in line with their actual rate of progress and therefore, they may not see a decrease in RIR from week to week with small increases. So, essentially, if someone was following a linear progression programme, where their goal is to increase load on the bar weekly by X kg, then they could make the load increase with that fixed RIR in mind and potentially make a smaller / larger load increase if the RIR target is in line with them doing so.

RPE/RIR & Training Planning – Addressing Common Concerns

1. time.
2. video feedback.
3. coaching feedback.
4. testing rep maxes vs training reps.

Relative Intensity FAQs

Before we conclude, I want to add in some frequently asked questions, as there are some additional questions worth answering that may not warrant their own section. If anyone happens to have further questions, I will happily add them to this section:

“If my time to train is limited (e.g. twice per week), would it be worth my while training to failure?”

In this case, keeping RIR lower and hitting failure on some sets would be appropriate. With the goal being to maximise the per-session training stimulus, working closer to failure is probably a good idea, especially considering you are probably doing full-body sessions. With at least 3 days of rest between those sessions, it’s going to be easier to come back and train again than if you were training 4+ times per week. However, having said that, you are still probably better served to save 0-1 RIR sets for your last working sets on exercises like squats, deadlifts, bench presses and overhead presses, where there is more skill involved and generally more fatigue accumulating. The thing is, you don’t want your first set to compromise your other sets, so saving an “as many reps as possible” (AMRAP) set for your last working set could be a good idea here. Similarly, you also have to keep an eye on what exercises are next. For example, if you have to squat after you deadlift, it may be a good recommendation to not work to failure on deadlifts, as it will very likely compromise your squat performance. This is the beauty of programming, there are rarely black-and-white answers, but there are many smart ways to do things.

“When should someone work to failure?”

There are a couple of instances where simply doing as many reps as you can with a given load is likely a good idea:

1. At the end of a training cycle where you have a lighter week of training to follow, working to failure isn’t a bad idea, as you know you have more time to recover, along with the fact that it is nice to see where your performance is at as a gauge for future training cycles. 
2. In general, taking isolation exercises to failure, particularly for smaller body parts, isn’t a bad idea. When a joint/muscle is isolated, the overall fatigue accumulated from working to failure is going to be far lower, for multiple reasons, both psychological and physiological. 
3. If you are short on time, working to failure or using a technique like rest-pause may allow you to get more “effective” work close to failure, providing a more potent training stimulus with less overall volume. 
4. It does seem that working to failure, preferably with higher repetitions, increases local muscular endurance, which may potentiate the performance of more total volume in subsequent training cycles and/or have benefits for athletes looking to develop muscular endurance in a specific muscle group. 
5. Finally, as a “teaching” method, training to failure can help you to get a better feel for where failure really is, but also how to keep your composure and technique in check when working to that point.

“Is there any place for very low RPE training?”

Yes, there are two particular uses for training with lower relative intensity:

1. For athletes looking to develop muscular power and/or speed, some high-velocity training is going to be useful. In order to move at high velocities, the relative intensity must be low. Therefore, performing sets of 3 (just an example) of a given exercise at a high velocity, even if it’s your 10-12RM, can have beneficial effects on power, as you are working on increasing your rate of force development. 
2. For anyone looking to increase their skill on a given movement, low relative intensity training is going to be very beneficial. In order to promote motor learning, practice is your best friend. However, if that practice is always in a fatigued state (i.e. when you are working with high intensities or close to failure), you will be limited in terms of the amount of practice you can get in. What this may look like in practice is 1-2 harder training days (e.g. 3×8 and 3×5 on the squat at 2 RIR), with one lighter day in between focused on technique development (e.g. 5 x 3 at 5 RIR).