The gym mantra says you must train to failure for maximum growth. Grunting, shaking, barely completing the last rep. But the research tells a more nuanced story: training to failure works, but stopping 1-3 reps short works just as well, with significantly less cost.
A meta-analysis of 15 studies (Refalo et al. 2023) found no meaningful advantage for failure, and in trained lifters RIR 1–2 grew muscle just as well as failure (Refalo et al. 2024): training at RIR 1-3 builds practically the same muscle as training at RIR 0 (failure). An exploratory analysis (Robinson et al. 2024) suggests a little more growth the closer to failure a set ends. The key insight: the last 5 or so reps of any hard set are the "effective reps" that drive growth. Whether you do 8 reps at RIR 2 (10 possible) or 10 reps at RIR 0 (failure), you get roughly the same number of stimulating reps.
Systemic fatigue: Failure sets generate disproportionate fatigue compared to the extra 1-2 reps gained. This fatigue reduces performance on subsequent sets, subsequent exercises, and subsequent sessions. Over a week, the cumulative fatigue deficit can meaningfully reduce total training volume.
Longer recovery: Taking every set to failure slows recovery over the next 24–48 hours compared with the same volume and clearly more reps in reserve (Morán-Navarro et al. 2017). For muscles trained 2x/week (Schoenfeld 2016), this can compress the recovery window.
Injury risk: Form breaks down at true failure. The last rep of a set to failure on squat or bench press carries significantly higher injury risk than the same exercise at RIR 1-2.
CNS fatigue: Repeated failure training taxes the central nervous system, leading to cumulative fatigue that manifests as decreased motivation, poorer sleep, and reduced performance across all exercises.
Last set of isolation exercises: The injury risk is low (lateral raises, curls, leg extensions), and the systemic fatigue cost is minimal. Taking the final set of an isolation exercise to RIR 0 squeezes out maximum stimulus with minimal downside.
Technique testing: Occasionally (once every 2-4 weeks), take a set to true failure to calibrate your RIR estimation. This improves your autoregulation accuracy without the chronic fatigue of daily failure training.
Deload week context: During a deload week with reduced volume, 1-2 failure sets won't create recovery problems because total volume is already low.
Heavy compounds (squat, bench, deadlift): Form degradation at failure on these exercises creates serious injury risk. Always keep RIR 1-2 on heavy compounds.
First sets of an exercise: Failing on set 1 of 4 devastates performance on sets 2-4. Better: RIR 2 on early sets, RIR 1 on the last.
During a calorie deficit: Recovery is already compromised. Adding failure training on top of a deficit is a recipe for overreaching.
Failure training frequency: Even for the exercises where failure is acceptable (last set of isolation), do not train to failure every session. Reserve true failure for 2-3 exercises per week. The rest should end at RIR 1-2. This gives you the stimulus benefit of occasional failure without the chronic fatigue accumulation of daily failure training across all exercises.
No. Studies show RIR 1-3 builds practically the same muscle (Refalo et al. 2023, 2024). Training to failure adds minimal extra stimulus with disproportionate fatigue cost. Reserve failure for the last set of isolation exercises.
Calibrate by occasionally taking a set to failure (every 2-4 weeks). Compare your estimated RIR with reality. Most beginners overestimate by 2-3 reps; this improves with practice. Our RPE vs RIR guide covers estimation in detail.
No. The effective reps in a set at RIR 2 are nearly identical to a set at RIR 0. What you lose: 1-2 extra stimulating reps. What you gain: less fatigue, higher volume capacity, lower injury risk, and better long-term progress.
MUSCLE TECHNICS prescribes an RIR target for every set: each set ends at least one rep short of failure. Maximum stimulus, minimum unnecessary fatigue. Science-based intensity without guesswork.
Try free for 14 days →