RIR vs Training to Failure: What the Data Says

Failure is not one switch. A series of meta-regressions on proximity to failure found hypertrophy increased as sets were taken closer to failure, while strength gains stayed similar across a wide range of RIR [1]. Two different curves for two different goals.

The authors themselves stress the analysis is exploratory — RIR was estimated by participants, not measured, so exact thresholds remain uncertain [1]. Treat the result as a direction: closer to failure probably buys more size, while for strength the variable barely moves the needle.

It helps to understand why those two curves diverge. Hypertrophy depends largely on how many reps in a set happen under full motor unit recruitment — and the last reps before failure meet that condition best. Maximal strength additionally depends on neural and technical factors, which are trained by load rather than by exhaustion.

So a set of 5 at 3 RIR can be an excellent strength stimulus and a mediocre hypertrophy one at the same time. That is not a contradiction in the data; it is two different goals measured in one set.

The cost of failure is not distributed evenly across exercises. A set of squats taken to technical failure generates systemic fatigue that spills into the rest of the session and into the following days. A set of curls to failure costs almost purely locally. Since strength is stable across a wide range of proximity to failure anyway [1], adding risk on the compounds buys very little.

A sensible default for intermediates: 1–3 RIR on squats, deadlifts and presses; 0–1 RIR on isolations at the end of the session. You keep rep quality where technique governs safety and spend proximity to failure where it is cheap.

Load7 enforces that split at the plan level: if your compounds start regularly landing at 0 RIR with missed reps, it drops the load rather than waiting for the week to fall apart.

The cost distribution also shifts within a session. The same set of squats performed first or fourth in the workout generates different residual fatigue, because in the second case you start it already in deficit. So proximity to failure is best planned on a descending scale: most conservative early on the compounds, most aggressive late on isolations.

A simple control: the last exercise of a session can go to 0–1 RIR, because nothing after it is left to ruin. The first exercise should leave a reserve, because everything that follows depends on its quality.

Most lifters overestimate how close to failure they are, especially on multi-joint lifts and early in a mesocycle. Calibration is simple: every few weeks take one isolation set to genuine failure and compare how many reps you got beyond your stated RIR.

The second method is an objective check. If you call 2 RIR but the last rep of the set shows no visible slowdown, you probably had four. You do not need a sensor for this — filming the set on your phone is enough to see the difference.

Calibration has one more dimension that rarely gets discussed: RIR drifts across a mesocycle. In week one, rested, you usually underestimate how many reps are left. In week six, with accumulated fatigue, the same set feels closer to failure than it really is.

That drift explains a familiar pattern: loads climb for three weeks and then stall, even though nothing changed in the plan. What changed is the calibration of your perception, so in real terms you are training with a growing — or shrinking — reserve depending on which way you drift.