RPE vs RIR: the number that tells you exactly how hard to push — backed by 4 RCTs

Two numbers that replace guesswork in the gym

You don't need to go to failure every set to keep making progress. You just need to know how close you are to it.

That's what RPE and RIR do. RPE — Rating of Perceived Exertion — is a 1–10 scale for how hard a set feels. A 10 is true failure, nothing left. An RPE 7 means you could still do 3 more reps. RIR — Repetitions In Reserve — is the flip side of that same coin. It's the raw count of reps you have left in the tank before you physically can't continue. RPE 7 = RIR 3. RPE 8 = RIR 2. They describe the same thing from opposite directions.

Most lifters still program by percentage of 1RM — your one-rep max — which is the most weight you can lift for a single rep. The problem: your 80% day after a bad night's sleep feels nothing like your 80% day when you're fresh. A fixed percentage is blind to how your body feels right now. RPE and RIR are not. They adjust in real time, automatically.

RIR-based training matches — or beats — going all-out

The first worry most people have: "If I don't push to failure, am I leaving gains on the table?"

Short answer: no — at least not for strength and power.

An RCT — people randomly split into groups and then compared — put 14 under-17 female basketball players into two 8-week programs. One group trained to failure (maximum-effort, or RM). The other trained with 3 reps left in reserve each set, which maps to an RPE of 7 on the 1–10 scale (Arede et al., 2020).

Both groups improved on every test. But the RIR group improved bench-press 1RM significantly more, with a large effect size (η²p = 0.40). Sprint, jump, and agility scores improved in both groups at similar rates. The authors concluded that leaving reps in reserve is a smart in-season strategy — it drives adaptation without the recovery cost of constant failure training (Arede et al., 2020).

That recovery cost matters. Grinding to failure every set spikes fatigue faster. When fatigue accumulates, you lose reps in later sets, which means less total volume — and how many sets per workout is what actually predicts muscle growth.

RIR-based training produced significantly greater bench-press strength gains in 8 weeks — without a single set to failure.

— Arede et al. (2020). Repetitions in reserve vs. maximum effort resistance training programs in youth female athletes. J Sports Med Phys Fitness.

The accuracy problem: RIR works better when you're experienced

Here's the catch: RIR only works if you can actually feel how many reps you have left.

Researchers tested this directly. Twenty-seven college-aged men — 14 experienced benchers (averaging 4.7 years of training) and 13 novices (averaging 1.1 years) — performed bench press sets at 60%, 75%, and 90% of their 1RM, then reported their RPE based on RIR (Ormsbee et al., 2019).

Experienced benchers were significantly more accurate. At their 1RM, experienced benchers reported an average RPE of 9.86 — almost exactly 10, which is correct. Novice benchers underestimated the same load. They also moved the bar faster at max effort, suggesting they hadn't yet developed the body awareness — the internal feedback loop — needed to sense proximity to failure (Ormsbee et al., 2019).

A second study confirmed this with a different population: older adults (average age 68) working at 65% of their 1RM. When asked to stop at RIR 2 (2 reps left), they actually stopped with significantly more reps left than they thought — underestimating by an average of 2.1 reps. At RIR 4, they underestimated by 1.6 reps. Only at RIR 6 — a very easy stopping point — were they accurate (Gómez-Redondo et al., 2025).

The closer you are to failure, the harder it is to judge — until you've logged enough sets to know what failure actually feels like.

RPE also changes by body part — not just by load

One more variable that trips people up: the same load doesn't feel equally hard on every exercise.

A study on 29 novice lifters running a 10-rep-max test found that RPE/RIR was significantly higher for lower-body exercises than upper-body exercises at every load from 50% to 100% of max (Cavarretta et al., 2022). In plain terms: a squat at 80% of your max feels harder than a bench press at 80% of your max — even though the percentage is identical.

Free weights also felt harder than machines at the same relative load. This is likely because free-weight exercises — like a barbell squat or a bench press — require more muscle coordination to stabilize the bar, which adds to overall perceived effort without necessarily adding more stimulus to the target muscle (Cavarretta et al., 2022).

Practical takeaway: don't use a flat RPE 8 across every exercise in your session and assume the effort is equivalent. Your RPE for a deadlift at 80% and your RPE for a cable curl at 80% describe very different biological situations.

Lower-body exercises scored significantly higher RPE than upper-body exercises at every load tested — the same percentage is not the same effort.

— Cavarretta et al. (2022). The Effects of Increasing Training Load on Affect and Perceived Exertion. J Strength Cond Res.

How to use RPE and RIR in practice

Here's the simple version of how to put this to work:

If you're new (under 1 year of consistent training): Use RIR 4–6. That means stopping when you feel like you have 4 to 6 reps left. It sounds conservative, but research shows you'll actually be closer to failure than you think — and being conservative protects you from the accuracy gap that novices have (Gómez-Redondo et al., 2025). As your body awareness sharpens, you can tighten that window.

If you're intermediate or experienced: Use RIR 2–3 (RPE 7–8) as your working zone for most sets. This keeps you close enough to failure to trigger adaptation — the physical changes that make you stronger — without the recovery cost of grinding to zero every set (Arede et al., 2020). Save RIR 0–1 for a final set, or for a testing day.

Adjust by exercise: Your RPE will naturally read higher on big compound lifts — squats, deadlift, rows — than on isolation work. That's expected. Don't force every exercise into the same RPE target.

Adjust weekly: If your RPE on the same weight is trending down session to session, that's your cue to add load. That's progressive overload training in action — and it's the principle RIR/RPE tools are built to serve.