Proximity to Failure: the Decisive Parameter
Introduction
It's the variable that most determines whether a set works, and the one worst controlled. Two people can do the same exercise, the same weight and the same reps, and be delivering completely different stimuli depending on how close to the limit they finished.
It's also where two opposite errors meet: those who never get close enough and those who go to failure on everything. Both pay a cost.
The reps that count are the last ones. The problem is that almost nobody knows how many they had left.
What it means
Muscular failure is being unable to complete one more rep with good technique. From there you count backwards with reps in reserve (RIR): 2 RIR means you could have done two more.
The relevant programming question isn't whether to reach failure, but how far from it to work.
What the evidence says
The picture has cleared up considerably in recent years, and not in the direction gym culture assumed.
For hypertrophy, Refalo and colleagues' meta-analysis found that getting closer to failure produces slightly more hypertrophy, with a small effect. Training 2 or 3 reps from failure produces results very close to training to failure, with considerably less fatigue.
For strength, the relationship inverts: reaching failure offers no advantage and in several studies is counterproductive. Strength improves by practising the movement with quality, and failure degrades technique and speed precisely on the reps executed worst.
For fatigue, the cost of failure is disproportionate. It increases recovery time between sessions and reduces performance in subsequent sets, so failing early in a session costs you effective volume later.
The practical conclusion is fairly clear: 1 to 3 reps in reserve covers nearly every case.
The estimation problem
There's a large obstacle between that recommendation and applying it: people estimate their reserve very badly.
Halperin and colleagues' review found that subjects systematically underestimate how many reps they could have done. It's common to believe you're 1 or 2 from failure when five remained. The bias is larger in less experienced people, in sets with light loads, and in leg exercises.
This means "I train at 2 RIR" is probably, in practice, 4 or 5 RIR. And there stimulus genuinely is lost.
How to calibrate: every so often, on a safe exercise — a machine, an isolation movement — take one set to true failure and compare with your estimate. It's uncomfortable, it's done a couple of times per block, and it corrects the bias considerably.
Where failure does belong
It isn't that failure is forbidden. It has its place:
In isolation exercises and machines, where technical risk is low and so is systemic fatigue.
On the last set of an exercise, when you won't need to perform in that pattern afterward.
With light loads, where approaching failure is mandatory for the set to stimulate anything.
Rarely in heavy basics. Failing a squat or deadlift with heavy load is technically and logistically bad, and adds nothing.
How to apply it by exercise type
- Heavy basics (squat, deadlift, bench): 2 to 3 RIR. Failure, essentially never.
- Compound accessories (leg press, rows, dumbbell presses): 1 to 2 RIR.
- Isolation (biceps, lateral raises, hamstring curl, calf): 0 to 1 RIR, occasional failure is fine.
And a rule that prevents the most common error: if you have four reps to spare on your last set, that set wasn't doing its job.
Common mistakes
Confusing muscular failure with technical failure. Continuing with your back or with a bounce isn't reaching failure: it's changing exercise mid-set.
Going to failure on everything. It generates a lot of fatigue, degrades technique and reduces the volume you can sustain across the week.
Never getting close. The opposite error, and probably more common among people training alone.
Estimating RIR without ever having calibrated. The number you record is a belief, not a measurement.
Using burn or shaking as the signal. They don't indicate proximity to failure; they indicate something else.
A note on safety
Taking sets to failure with heavy loads and no safety pins or spotter is the situation where most gym accidents happen. If you're going to approach your limit on bench press or squat, use the rack's safety pins or have someone spot you. With machines and isolation the risk is lower.
Conclusion
Proximity to failure decides whether a set stimulates anything, and the right range for almost all work is one to three reps in reserve. Closer adds little and costs a lot of fatigue; further away adds nothing.
Do the calibration this week: on a machine exercise, estimate your RIR on the last set, write it down, and continue to true failure to see how many were actually left. You'll likely discover you've spent months training further from the limit than you thought, and that single adjustment is worth more than a change of routine.
References
- Refalo, M. C., et al. (2023). Influence of resistance training proximity-to-failure on skeletal muscle hypertrophy: A systematic review with meta-analysis. Sports Medicine, 53(3).
- Grgic, J., et al. (2022). Effects of resistance training performed to repetition failure or non-failure on muscular strength and hypertrophy: A systematic review and meta-analysis. Journal of Sport and Health Science, 11(2).
- Halperin, I., et al. (2022). Accuracy in predicting repetitions to task failure in resistance exercise: A scoping review. Sports Medicine, 52(2).
- Zourdos, M. C., et al. (2016). Novel resistance training-specific rating of perceived exertion scale measuring repetitions in reserve. Journal of Strength and Conditioning Research, 30(1).
- Vieira, A. F., et al. (2021). Session rating of perceived exertion following resistance exercise with different loads. International Journal of Environmental Research and Public Health, 18(11).