Key takeaways
- Momentary failure means you cannot complete another rep with the technique you started with, not that the set felt hard.
- When total work is matched, stopping short of failure produces comparable muscle growth to going all the way.
- Strength gain appears to be the variable least sensitive to how close you train to failure.
- Failure sets leave measurably more neuromuscular fatigue behind, which costs you sessions later in the week.
- Reps in reserve is the practical control: 2-3 RIR on heavy compounds, 0-1 RIR on isolation.
Failure is the most argued-about variable in lifting, and one of the least useful ones to argue about. The camp that insists every set must end in mechanical collapse and the camp that treats failure as always counterproductive are both describing edge cases. What the evidence points at is less dramatic and considerably more useful: for most sets, in most weeks, stopping a rep or three before the point of collapse buys you the growth without most of the bill.
What failure actually means
Momentary muscular failure has a specific definition: the point at which you cannot complete another concentric repetition with the technique you started the set with. That last clause does most of the work. A set where the bar path collapses, the hips shoot up, or a training partner starts lifting for you did not reach muscular failure. It reached technical failure several reps earlier, and everything after that trained something other than the muscle you were aiming at. Where a set should stop is a programming decision, and every program in the Method library states it as reps in reserve rather than leaving it to feel.
Failure is also not the same as discomfort. The burn arrives well before the muscle runs out of force, which is why people routinely believe they hit failure when a video of the set shows the bar still moving quickly. The gap between "this is unpleasant" and "this rep will not complete" is wider than almost anyone estimates from the inside of a set.
That gap is why coaches use reps in reserve, or RIR: the number of additional reps you believe you could have completed. Zero RIR is failure. Two RIR means you had two left. It is an estimate, not a measurement, but it is a far more useful instruction than "go hard."
What the research shows for muscle growth
This question has been examined directly. A systematic review with meta-analysis pooling the studies that compared training closer to versus further from failure found no meaningful hypertrophy advantage for taking sets all the way, once the other variables were held constant (Refalo et al., Sports Med 2023). Sets stopped short of failure built muscle comparably to sets that went to it.
A later series of meta-regressions modeled the relationship as a curve rather than a yes-or-no comparison, and found a small trend favoring closer proximity to failure for hypertrophy, with strength gain largely unaffected by where sets stopped (Robinson et al., Sports Med 2024). Read together, the picture is not "failure does nothing." It is that the effect is modest, and it is not the lever most lifters think it is.
The lever that does move consistently is how much hard work you accumulate. The dose-response relationship between weekly set volume and muscle growth is one of the better-replicated findings in the training literature: more hard sets per muscle per week, up to a point, produced more growth (Schoenfeld et al., J Sports Sci 2017). Which reframes the whole question. Failure is not free volume. It is a way of spending the volume you can recover from, and spending it faster.
Strength is a different question
Strength adaptation is substantially a skill adaptation. You are teaching a movement pattern to express force efficiently, and grinding reps teach the pattern badly. The meta-regression work found strength gain to be the outcome least sensitive to proximity to failure (Robinson et al., Sports Med 2024), which lines up with what strength coaches have argued for decades: the heaviest, most technical lifts should be performed crisply, with reps in the bank.
Practically, that means a top set of squats or bench press should end while the bar is still moving with intent. If rep speed has dropped off a cliff, the set is over regardless of whether one more could theoretically be forced.
The cost nobody budgets for
The argument against routine failure training is not that it fails to stimulate. It is that it charges more for the same stimulus. When the same exercises were performed either to failure or stopped short, the failure condition produced longer-lasting neuromuscular fatigue and a slower return to baseline performance (Morán-Navarro et al., Eur J Appl Physiol 2017). That fatigue does not stay inside the session that generated it.
If you train four or six days a week, every set taken to failure on Monday is a small withdrawal from Wednesday and Friday. The lifter who takes every set to the wall usually ends up training fewer effective hard sets across the week than the lifter who stops two short, because the second lifter can actually complete the program.
Reps in reserve, in practice
RIR is only useful if your estimate is roughly calibrated. Three habits fix most of the drift:
- Watch bar speed, not the burn. The rep at which speed noticeably slows is usually somewhere around 2-3 RIR on a compound lift. Discomfort is a poor proxy; velocity is a decent one.
- Recalibrate occasionally. Every few weeks, take one set of a safe machine or cable movement to genuine failure. Compare it to what you assumed 0 RIR felt like. Most people discover they have been stopping earlier than they thought.
- Write the number down. Logging "3 x 8 @ 2 RIR" turns effort into something you can progress deliberately, the same way you progress load.
Once effort is a number, it becomes another lever for progressive overload: the same weight for the same reps at a lower RIR is real progress, even before the load changes.
Where failure earns its place
Failure is not banned. It is a tool with a specific bill attached, and some movements have a much smaller bill than others.
| Movement type | Where to stop | Why |
|---|---|---|
| Heavy barbell compounds squat, deadlift, bench, row | 2-3 RIR | Technique degrades before the muscle does, and the systemic fatigue cost is highest here |
| Machine and cable compounds leg press, chest press, pulldown | 1-2 RIR | Stable and self-limiting, so you can push closer without the pattern falling apart |
| Single-joint isolation curls, lateral raises, leg curls | 0-1 RIR | Low systemic cost, quick to recover, and hardest to load meaningfully any other way |
| Final set of the session | 0 RIR, occasionally | Nothing is left to compromise, so it is the cheapest place to spend a failure set |
How to set it across a training week
Proximity to failure and training frequency trade against each other. The more often you train a muscle, the further from failure each set should sit, because you need to be functional again in forty-eight hours. The less often you train it, the more you can afford to push, because there is a longer runway before the next session.
On a six-day Push / Pull / Legs rotation, where everything gets hit twice a week, 2-3 RIR on compounds is the setting that keeps the week intact. On a four-day Upper / Lower split, the same logic applies with slightly more room on accessories. On a three-day Full Body plan, with more recovery between sessions and fewer sets per muscle per week, you can reasonably run the last set of each movement closer to the line. Every plan in the OPTML Method training library specifies this per set, so the decision is not left to how you feel that day.
An attribution worth getting right: the MV / MEV / MAV / MRV volume-landmark language you see everywhere online is Renaissance Periodization's framework, developed by Mike Israetel. It is a coaching model for organizing weekly set counts, not a trial result. Schoenfeld's dose-response work is the published evidence for the general shape those landmarks describe. The two get blurred constantly. They are not the same kind of claim.
The mistakes that make it backfire
Failing on the first set. Take set one to the wall and sets two through four collapse, so the session ends with less total quality work than if you had stopped short on all four.
Failing on the lifts that punish it. Deadlifts and squats to failure buy a rounded back and three days of soreness. The stimulus is not worth the exposure.
Calling form failure "failure." If the last three reps looked nothing like the first three, you trained your hip flexors and your ego.
Never approaching failure at all. The opposite error is just as common. Sets ended at 5 RIR because they started to feel like work do not accumulate meaningful stimulus. "Short of failure" means one to three reps short, not five to eight.
Using failure to paper over low volume. Two brutal sets do not substitute for the weekly set count the evidence actually points to (Schoenfeld et al., J Sports Sci 2017). Add sets before you add suffering.
Bottom line
Train most sets one to three reps shy of failure, push closer on isolation and machine work, keep the heavy barbell lifts crisp, and let total weekly volume do the heavy lifting. The evidence does not support the idea that every set must be an all-out effort, and it does show that all-out efforts cost more recovery than the sets that stop short. The lifters who make the most progress over years are rarely the ones training hardest in any single set. They are the ones whose effort is calibrated well enough that they can repeat it on Wednesday.
Educational content, not medical advice. Individual results vary.
