Do You Need to Train to Muscular Failure? RIR and Muscle Growth

Momentary muscular failure is the point at which, despite maximal effort, you cannot complete another repetition with proper technique. It is not the same as lack of motivation or pain; it is a temporary inability of the muscle to generate enough force. The literature also uses broader definitions of “set failure”, for instance when movement speed clearly drops or a coach decides to end the set, which is why results of different papers cannot always be compared directly.

RIR (repetitions in reserve) is the number of repetitions you estimate you could still have done before failure. A set ended at 0 RIR means failure; 2 RIR means you stopped with about two reps left. In the popular scale described by Zourdos et al. (2016), a rating of perceived exertion is linked to RIR (RPE 10 = 0 RIR, RPE 9 = 1 RIR and so on), and a study of 29 people found a strong inverse relationship between movement velocity and perceived exertion in both experienced and novice lifters. The authors considered the scale a practical way to regulate training load.

Why would proximity to failure matter at all? The leading mechanistic hypothesis holds that for muscle growth it is key to recruit a large number of motor units under sufficiently high mechanical tension; with light loads or early reps in a set, some fibres are not yet recruited, so only the last, difficult repetitions are “effective”. The explanation is plausible but remains a hypothesis; it has not been directly measured in humans at the level of every motor unit, so we base conclusions on longitudinal training studies.

An effect size of 0.19 with the lower confidence limit exactly at zero means that the difference, if it exists, is very small. It is also worth noting how important the definition is: when momentary failure was strictly compared with non-failure, the result was statistically indistinguishable from zero. That subgroup analysis is the most relevant to the question “must I push every set to the end?”.

The notion of a “non-linear relationship” is practically key. It suggests the first steps towards more effort (from 6 RIR to 3 RIR) give a lot, while the last ones (2 RIR to 0 RIR) give little but cost disproportionately more fatigue. The authors cautiously call this a potential relationship, not an established fact.

Both meta-analyses agree on the main message: failure is not a necessary condition. The small advantage in trained individuals in this paper (ES 0.15) may suggest that the more advanced the lifter, the more final effort matters — consistent with the intuition that a beginner needs almost any stimulus while an advanced lifter needs more. We do not treat this as final proof, since it is a subgroup result.

For maximal strength the picture is even simpler: failure does not appear necessary, and when volume was not equated, non-failure training came out ahead. The practical takeaway for people training mainly for strength is that heavy sets should be ended with some reserve — quality and speed of repetitions matter more than grinding to the end, especially in main lifts.

This study seems to contradict the earlier ones, but on close reading it complements them. Categorical meta-analyses (failure versus no failure) compare the point at RIR 0 with a broad, heterogeneous set of “non-failure” — from RIR 1 to RIR 6 or more. A meta-regression looks at the whole continuous range and sees that the further from failure, the weaker the hypertrophy. Reconciling both gives a practical conclusion: you do not need to reach RIR 0, but it is not worth stopping at a comfortable distance of 5–6 reps from the limit either.

This study is valuable because it uses a within-participant design: each participant had one leg trained to failure and the other with reserve, which removes genetic and lifestyle differences. It has a small sample (18 people), covers two leg exercises and 8 weeks, so it is an illustration, not a resolution. It does show a key mechanism: in sets to failure there is a larger loss of repetitions and velocity, which affects the volume of subsequent sets; ending at 1–2 RIR lets you maintain higher work quality.

This result has major practical significance. If the average lifter is off by 2–3 repetitions (typically on the low side, meaning they think they have less in reserve than they truly do), then “3 RIR” in their perception may in practice be 5–6 RIR. This means people convinced they work close to failure are often much further from it. That is all the more reason to occasionally — in safe exercises — actually reach failure in order to calibrate your feel.

Accuracy of RIR estimates also depends on rep range, exercise and fatigue, and in studies it generally improves for sets closer to failure and in more experienced people. Simple calibration methods include filming sets and watching for slowing movement speed, noting how many repetitions you actually did in a later set to failure, and comparing with previous weeks.

If the benefit of failure is at most small, we should ask about costs. The first is neuromuscular fatigue and lost volume: in the study in the previous section, loss of velocity and repetitions was consistently greater with training to failure. If after a set to failure you can do far fewer repetitions in the next set, total work in the session falls, and that is one of the main variables influencing growth (we write about weekly volume in the article on sets per muscle group).

The second cost is longer recovery. A set to failure typically causes more fatigue and potentially more muscle damage, which can affect the quality of subsequent workouts for the same muscle; in practice many people feel this as longer soreness and a drop in strength the next day. The evidence in the literature comes from short studies, though, and it is hard to give one number for “how many days longer”. We treat this as a reasonable assumption, not a firm result.

The third cost is safety. In free-weight exercises where failure means losing control of the load (barbell squat, bench press, overhead press), a set to failure without spotting carries a real injury risk. In machine and cable exercises, where you can safely stop the movement, the risk is much lower. Add to this something that is hard to capture in studies: the last reps at the edge of failure are often performed with worse technique, which raises load on joints and tendons.

The ranges above are not the result of one study but a synthesis: of meta-analyses showing failure is not required, of a meta-regression suggesting benefit of proximity to failure, of caution stemming from errors in RIR estimation, and of common sense about safety. Adapt them to your experience and the exercise you are performing.

Likewise, the myth “no burn, no effect” confuses sensation with stimulus. The burning feeling in the muscle results from metabolite accumulation and is not a reliable indicator of whether a set was close enough to failure. A better gauge is the number of extra repetitions you could have done or the drop in movement speed.

If we had to put forward one rule, it would be: work close to failure, but you do not need to touch it. End most sets with 1–3 reps in reserve and leave failure for the last set of safe isolation exercises. That way you keep high quality and volume of work, limit fatigue and injury risk, and still do not lose stimulus.

The biggest practical lesson of these studies is not failure itself, though, but how often we underestimate — or overestimate — it. If you have never tested your limit in safe conditions, your “3 in reserve” may really be “6 in reserve”, and if you grind every set to the end every day, you are probably paying for it with fatigue that limits volume. Calibrating your own feel is cheaper than adding more sets.