Training to Failure: Science behind it

A lifter grinds out a final rep, bar speed collapsing, form barely holding together. Two racks over, someone stops the same exercise with what looks like gas left in the tank. Both are following legitimate advice. The internet has held strong opinions on both sides of this question for years. The research, taken as a whole, gives a more precise answer than either camp typically offers. This article covers what training to failure does for strength and hypertrophy. It covers where the two outcomes diverge, and what that means for how close to failure a set should go.

---

What “Training to Failure” Means

Training to failure means continuing a set until another repetition cannot be completed with proper form. Researchers quantify how far a set stops short of this point using repetitions in reserve, or RIR. This is the number of additional reps a lifter believes they could still perform. A set stopped at 2 RIR means the lifter judged two more reps were available before technical failure.

This measurement matters because most research does not simply compare “failure” against “not failure.” It compares specific RIR targets against each other. That allows a more precise picture of what happens as a set gets closer to, or further from, its true endpoint.

---

---

The Central Finding: Strength and Hypertrophy Diverge

The most methodologically rigorous analysis on this question is a 2024 meta-regression. It treats proximity to failure as a continuous variable, rather than a simple binary. The results split cleanly along two different outcomes. For strength gains, the confidence intervals across the full range of RIR values contained a null result. Training closer to failure showed no meaningful advantage over stopping several reps short. For muscle hypertrophy, the relationship moved in the opposite direction. Sets terminated closer to failure produced measurably greater increases in muscle size, a relationship that held up statistically across the pooled data.

This single finding reframes the entire debate. Training to failure and training with reps in reserve are not simply better or worse versions of the same thing. They appear to be somewhat different tools. One shows a real, if modest, hypertrophy advantage. Neither shows a clear strength advantage from pushing closer to true failure.

---

How Big Is the Hypertrophy Advantage, Really

The advantage exists, but it is worth being precise about its size rather than treating it as a mandate. A 2023 meta-analysis pooling 15 studies found a small effect favoring training closer to failure for hypertrophy outcomes. The effect size sat in the range of 0.15 to 0.20, a modest advantage by any standard.

A controlled trial randomized 26 resistance-trained adults to either true failure or 1 to 2 RIR across 8 weeks. Quadriceps hypertrophy was similar between both groups. A separate 8-week study compared failure training against 2 RIR training, using single-set routines across nine exercises. Several hypertrophy measures tended to favor training to failure, though the absolute differences were generally modest. The pattern across this body of research is consistent. Stopping 1 to 3 reps short of true failure appears to capture nearly all of the achievable muscle growth. The final push to true failure adds a small amount more, at a real cost worth understanding before chasing it as a default strategy.

---

---

Why Strength Doesn’t Follow the Same Pattern

The strength finding deserves its own explanation rather than being read as a simple absence of benefit. One study directly compared resistance-trained adults across 5 weeks of squat, bench press, and deadlift training. A low-RIR group trained at 0 to 1 RIR; a high-RIR group trained at 4 to 6 RIR. The low-RIR group showed changes in motor unit firing characteristics consistent with increased recruitment of lower-threshold motor units, a genuine neuromuscular adaptation.

This suggests training closer to failure does produce real physiological changes. Those changes simply do not appear to translate into additional strength beyond what moderate-RIR training already achieves. A plausible explanation is that strength adaptation depends heavily on practicing the specific skill of producing maximal force. That happens effectively across a wide range of proximity to failure. The additional neuromuscular fatigue accumulated at true failure does not add a proportional return.

---

The Fatigue Cost Is the Part Worth Weighing Most Carefully

Training to failure produces measurably more acute fatigue than submaximal-effort training, a finding well established in the broader fatigue literature. This cost matters because training does not happen in isolation. A single set taken to failure affects how much quality work follows it. This holds within the same session and across the following days.

This is where the practical calculation resolves. If failure training added a large, unambiguous hypertrophy advantage, the fatigue cost would be an easy trade to accept. The advantage is small, roughly a 0.15 to 0.20 effect size according to current meta-analytic evidence, and strength shows no advantage at all. That makes the fatigue cost the deciding factor for most lifters most of the time. Spending that recovery capacity on additional volume, rather than on pushing existing sets closer to failure, is frequently the better trade across a full training week.

---

Where the Evidence Is Thinner

A few caveats are worth stating plainly rather than glossing over. Research on proximity to failure is considerably denser for strength and hypertrophy than for other outcomes. Power and explosive performance have not been mapped with the same density of evidence. What has been published tends to suggest staying well clear of failure for power-focused work, though the available sample of studies is smaller and less conclusive.

RIR estimation accuracy also varies by exercise and by lifter. One study found participants were more accurate estimating reps in reserve on the bench press than on the squat. Accuracy improved with practice over the course of the intervention. A prescribed RIR target is only as useful as a lifter’s ability to judge it accurately in the moment. That judgment is itself a trainable skill, not a fixed trait.

---

A Practical Framework

For hypertrophy-focused training, stopping most sets at 1 to 3 RIR captures nearly all of the available muscle growth. Occasional sets can be pushed closer to or at failure on isolation exercises specifically, where the fatigue cost and injury risk are lower than on heavy compound lifts. For strength-focused training, there is little evidence supporting a need to train closer to failure than 2 to 4 RIR. Doing so consistently risks accumulating fatigue that undermines subsequent training volume and quality, without a corresponding strength return.

The honest summary is that training to failure is a legitimate tool, not a myth to be dismissed. It has a narrow, specific use case, not a default setting. It earns its place occasionally, on the right exercises, for the right goal, rather than as the standard endpoint of every set.

---

References

1. Robinson ZP, Pelland JC, Remmert JF, et al. Exploring the dose-response relationship between estimated resistance training proximity to failure, strength gain, and muscle hypertrophy: a series of meta-regressions. Sports Med. 2024. DOI: 10.1007/s40279-024-02069-2
2. Refalo MC, Helms ER, Trexler ET, Hamilton DL, Fyfe JJ. Similar muscle hypertrophy following eight weeks of resistance training to momentary muscular failure or with repetitions-in-reserve in resistance-trained individuals. J Sports Sci. 2024;42(1):85-101. DOI: 10.1080/02640414.2024.2321021
3. Refalo MC, Helms ER, Trexler ET, Hamilton DL, Fyfe JJ. Influence of resistance training proximity-to-failure on skeletal muscle hypertrophy: a systematic review with meta-analysis. Sports Med. 2023;53(3):649-665.
4. Effects of resistance training to near failure on strength, hypertrophy, and motor unit adaptations in previously trained adults. Physiol Rep. 2023. PMC10161210.
5. Without fail: muscular adaptations in single-set resistance training performed to failure or with repetitions-in-reserve. Med Sci Sports Exerc. 2025. PMID: 40249908