Stopping 1-3 reps short matched training to failure over eight weeks
In short
Nineteen trained participants (11 men, 8 women, mean age 22.3) spent eight weeks doing leg extensions with one limb taken to momentary muscular failure and the other stopped at 1-3 repetitions in reserve. Bayesian mixed-effect models estimated a between-protocol difference of -0.019 cm (-0.094 to 0.054) in knee-extensor thickness and 0.198 kg (-1.87 to 2.6) in 1RM, both intervals spanning zero. Under these conditions there was no evidence that going to failure beat leaving 1-3 reps in reserve.
Whether to grind out the last rep is a decision you make every set. An eight-week trial priced it at 0.019 cm of muscle thickness and 0.198 kg of 1RM. Neither is something you could measure on yourself or load on a bar.
Why compare two legs of the same person?
Nineteen trained participants took part (11 men, 8 women, mean age 22.3 years, mean body mass 69.9 kg). One limb was randomly assigned to the failure (FAIL) protocol and the contralateral limb to a 1-3 RIR protocol. The exercise was a single unilateral leg extension.
Split people into groups and genetics, sleep, food and stress differ between them. Split limbs within a person and all of that applies equally to both sides. The only thing left to differ is where the set ended.
How close were the results?
Knee-extensor thickness was measured by ultrasound and 1RM tested before and after. The estimated between-protocol difference was -0.019 cm (-0.094 to 0.054) for thickness and 0.198 kg (-1.87 to 2.6) for 1RM. Both intervals comfortably span zero, and secondary standardised mean differences pointed to no clear advantage either way.
So why train to failure at all?
Failure buys certainty that the set was hard enough. What it costs is the question. If three reps in reserve produces the same adaptation, those three reps are better left in the recovery budget, which is exactly why the authors put RIR-based termination forward as a practical way to manage fatigue.
That matters most on heavy compound work. Take every squat or deadlift set to failure and the bill arrives in the exercises that follow and in the next session. A near-zero difference on a single-joint movement weakens the case for paying it.
Nineteen participants, one single-joint exercise, eight weeks. Nothing here guarantees the gap stays at zero for heavy multi-joint lifts or over longer blocks. And 1-3 RIR only works if you can actually judge how many reps are left.
What it means for your Muscle Index
Muscle Index is built from the sum of squat, bench and deadlift 1RM. If where you end a set is worth about 0.2 kg of 1RM after eight weeks, the variable that moves your score is how many sessions you completed intact, not whether each one ended in failure. Grinding to failure and losing the next session is a net loss on the index.
If judging reps in reserve feels like guesswork, read RIR is a trainable skill, not a feeling, and for a different way to end a set see cutting sets at 10% velocity loss.
Frequently asked questions
Do you have to train to failure to build muscle?
Not in this trial. Across eight weeks of leg extensions in 19 trained participants, the limb taken to failure and the limb stopped at 1-3 reps in reserve differed by 0.019 cm in knee-extensor thickness.
Was there any difference in strength gain?
No. The between-protocol difference in 1RM was 0.198 kg with an interval of -1.87 to 2.6 kg, which spans zero. Neither termination strategy could be called better.
What does RIR mean?
Repetitions In Reserve, meaning how many more reps you could have done when you ended the set. Training at 1-3 RIR means stopping while one to three reps are still available.
Why was the comparison made within the same person?
To hold genetics, sleep, diet and stress identical across the two conditions. When one person's limbs take different protocols, the set-termination point is the only variable left to explain the difference.
Does this apply to squats and deadlifts?
Not directly. The study used one single-joint exercise, 19 participants and eight weeks. It provides no evidence about heavy multi-joint lifts, where fatigue costs are larger.
Source: PubMed