How hard should a set be? This question sits at the centre of most effort management decisions in resistance training, and it's harder to answer than it appears. Percentage of one-repetition maximum (%1RM) is the traditional answer — expressed as a percentage of the heaviest single repetition you could perform. The problem is that 1RM is highly variable day to day, and the relationship between a given percentage and how it actually feels shifts with fatigue, training history, and the specific exercise.
Reps in Reserve (RIR) is a more practical way to express proximity to failure. It describes how many additional repetitions you could have completed at the end of a set if you had continued — a direct measure of effort relative to your actual capacity on that specific set.
What the evidence says about training near failure
Research on hypertrophy (muscle growth) consistently finds that sets taken close to failure — within 0–4 reps of it — produce similar hypertrophic outcomes to sets taken to absolute failure, with meaningfully less accumulated fatigue. Training to failure on every set is not necessary for growth, and carries costs in terms of recovery and cumulative fatigue.
The implication is that there is a productive range: close enough to failure to recruit high-threshold motor units and generate the tension required for adaptation, but not so close that every set is a maximal effort with significant recovery consequences.
That range is generally described as RIR 1–3: ending a set with 1 to 3 repetitions left in reserve.
How to calibrate your own scale
The most common error with RIR is underestimating it — believing a set is close to failure when it isn't. Most people, particularly newer lifters, stop further from failure than they perceive. A set that felt like RIR 2 was often RIR 5 or 6 by an objective measure.
The only reliable way to calibrate the scale is to occasionally take sets all the way to technical failure — the point where another full repetition cannot be completed with acceptable form — and count backwards from there. Doing this periodically, in a controlled way, gives you a reference point for what genuine proximity to failure feels like.
Once calibrated, RIR provides a more useful effort target than percentages, because it automatically adjusts for daily variation in capacity. A day where your 1RM is lower than usual — because you're slightly fatigued, or slept badly — will produce heavier sets if you're chasing a percentage; will produce appropriately adjusted sets if you're chasing an RIR target.
Where RIR sits in a training block
RIR targets typically decrease as a training block progresses. Early in a mesocycle, higher RIR (3–4) is appropriate — sessions are building the pattern, volume is accumulating, fatigue is being managed. As the block approaches its peak, RIR decreases (1–2), pushing closer to the maximal training stimulus. A deload week then follows, allowing fatigue to clear before the next block begins.
This progression is the structure beneath most well-designed programmes, even when RIR isn't explicitly named.