Somewhere around the seventh rep, a second count starts running. Not the set count on the program. The one underneath it. The quiet arithmetic that tracks how many reps are left in the body before the bar would stop moving on its own.
That number is a guess. It has always been a guess. Two reps left, maybe three, maybe one. The distance between the weight moving and the weight winning is felt, not measured, and the margin shifts from set to set.
Everyone who has ever racked a bar with reps still in them has made this call. The question of how many reps you should leave in the tank runs underneath every working set, and almost nobody has a grounded number behind their answer.
The fitness internet settled this into a binary years ago. One camp pushes every set to the brink. Failure or nothing. The other prescribes a generous buffer: three, four, five reps held in reserve to manage fatigue and protect recovery. Both positions sound like common sense because each one is partially right.
How Many Reps Should You Leave in the Tank
Training closer to failure produces more muscle growth, with one to three reps from failure capturing most of the growth benefit. Strength gains show no meaningful link to how close the set gets to failure, remaining similar whether a lifter stops one rep short or four reps short. The answer depends on the goal.
— Robinson et al. 2024 · Sports Medicine · meta-regression
The problem is that this was never a binary question. A 2024 meta-regression reframed the question. It treated reps in reserve as a sliding scale, not an on-off switch. It did not return one number. The answer split by goal.
When the outcome was muscle size, the signal was clear. The closer a set was taken toward the point of failure, the more the muscle grew. That was not a pattern that might have been noise. The signal was strong enough that chance alone could not explain it.
For strength, the picture was entirely different. Strength gains were similar whether a lifter stopped one rep short of failure or four reps short. The proximity that mattered so much for muscle size barely registered when the outcome was raw force.
The practical zone is narrow for one goal and wide for the other. For muscle growth, one to three reps from failure captures most of the benefit. Close enough to send the growth signal. Far enough to avoid the accumulated fatigue that consistent failure training stacks across a full week. For strength, the leash is longer. There is room to stop well short and still progress.
When the goal is muscle growth: One to three reps from failure captures most of the hypertrophy benefit.
When the goal is strength: Stopping well short of failure still produces similar strength gains.
But the map is not finished. The direction is established: closer to failure does produce more muscle growth. Whether two reps from failure is meaningfully different from one remains a question the evidence cannot answer with precision yet. No previous analysis had mapped proximity as a gradient rather than a switch, and it was explicitly flagged as exploratory. The data did not explain everything. Total sets, exercise selection, and individual recovery all share the stage. Proximity to failure is a real lever for muscle growth, though the program has other levers pulling alongside it.
When the weight is in your hands again and the internal count picks up, push or hold back is the wrong frame. What you came to build in that set decides the answer. If the session is about size, the answer lives closer to the edge: one to three reps from the limit. If it is about strength, the margin can be wider, and the body still adapts.
A question still sits past this answer: whether failure itself, the rep where the bar stops, is ever worth hitting on purpose. And whether training closer to that edge changes how many sets a muscle can recover from each week.