RPE and RIR Explained: How Hard Should Your Sets Be?

Published October 25, 2026

Collage of a lifter grinding out the last hard rep of a barbell bench press, face strained, with the headline "RPE & RIR Explained"

RPE (rating of perceived exertion) is a 1-10 score for how hard a set was, and RIR (reps in reserve) is how many more reps you could have done when you stopped. In lifting they measure the same thing: RPE 10 means no reps left, RPE 9 means one left, RPE 8 means two left. For most working sets, finish with about one to three reps in reserve (RPE 7-9), and expect, on average, to have about one more rep in the tank than you think.

One scale, two names

The version of RPE that lifters use is anchored to reps left, which makes it far more concrete than a vague "how hard did that feel." It was linked to bar speed in a 2016 study (Zourdos et al., Journal of Strength and Conditioning Research) of 29 squatters, 15 experienced and 14 novice. Reported RPE rose as average bar velocity fell, with a strong inverse relationship in both groups (r = -0.88 in the experienced lifters, -0.77 in the novices).

The conversion is simple:

RPE 10 = 0 RIR. You hit failure or could not have completed another rep.
RPE 9 = 1 RIR. One more rep was possible, but it would have been a grind.
RPE 8 = 2 RIR. The last rep was slow, yet you knew two more were there.
RPE 7 = 3 RIR. Hard work, with the bar still moving well.
RPE 6 = 4 RIR. Moderate. This is where guesses get unreliable (more below).
RPE 5 or lower = 5+ reps left. Warm-up territory.

Many lifters add half points, such as 8.5 for "one or two left." Guesses typically run about a rep off (more below), so treat half points as shorthand rather than precision. Two cautions apply. RPE describes effort, not load: a set of 12 at RPE 8 and a set of 5 at RPE 8 stress you differently, so always write the reps next to the number. And RPE 10 does not mean "it felt awful." It means you were out of reps.

How accurate is your guess?

Not perfect, and the error leans in one direction. A pooled analysis by Halperin and colleagues (Sports Medicine, published online in 2021) combined studies covering 414 participants and found people underpredicted their reps to failure by about 0.95 reps on average (95% confidence interval 0.17 to 1.73, with a lot of variation between studies). Predictions were more accurate when they were made closer to the end of the set, when sets went to fewer total reps (about 12 or fewer), and slightly more accurate in later sets. Training experience did not significantly explain the differences in that analysis.

Individual studies fill in the picture. In a 2024 study (Refalo et al., Journal of Strength and Conditioning Research), 24 resistance-trained men and women performed bench press sets to failure at 75% of their one-rep max and called out the rep where they thought they had one or three left. Their average miss was 0.65 reps in absolute terms, with a slight lean toward underpredicting. That is good accuracy for a trained group on a stable lift, and neither sex, years of training nor relative strength explained who did better.

A 2023 study (Remmert et al.) tested 58 trained and untrained men and women on machine-based biceps curls, triceps pushdowns and seated rows. Sex and experience again did not matter, but predictions were more accurate closer to failure and in later sets. Stronger By Science's summary of Remmert's work reports an average miss of about 1.2 reps at five in reserve, against roughly half a rep at one.

New lifters are a different story. In a 2017 PeerJ study (Steele et al.), 141 people took single sets to failure, mostly on machine exercises. The least experienced group underpredicted by roughly four to five reps, while the most experienced group missed by about one to two. The authors urged caution when prescribing training by RIR.

Most of these studies are modest in size, test a few exercises at set loads, and have a researcher watching. A real gym set, with fatigue and a half-remembered program, is messier. The practical takeaways:

If you are an experienced lifter and finish a set at what feels like 1-2 reps left, you are probably within about a rep.
When you think you have four or more left, treat it as a rough estimate. It can be off by several reps.
Sets of more than about 12 reps are harder to judge than shorter sets.
If in doubt, add one rep to your guess before deciding whether the set was hard enough.

How to calibrate your own RIR

Judging reps left is a skill that improves with practice. In a 2026 study (Wiedenmann et al., BMC Sports Science, Medicine and Rehabilitation), 13 younger adults (average age 33) and 13 older adults (average age 67) practiced estimating 4 and 2 reps in reserve on bench press and leg press across six sessions, then continued each set to failure. Their average error fell by up to 2.3 reps from the first session to the last, with no significant difference between age groups. The sample is tiny, but it points the right way: this is trainable.

A simple calibration routine:

Pick a lift where failure is safe: leg press with safety stops set, a chest press machine, leg extension, a cable row, or a barbell lift in a rack with safeties or a spotter.
After warming up, choose a weight you expect to lift for about 8-10 reps.
When you feel you have two reps left, say "two" out loud or note the rep number. Keep going until you cannot finish another rep with acceptable form, and count.
Compare. If you called "two" and got four more, your "two" is really closer to four.
Repeat over several weeks on different exercises, always as the last set of that exercise so the extra fatigue does not wreck the rest of your session.

Pay attention to cues as well: the bar slows noticeably, your form starts to shift, or your face tightens. Bar speed does track effort (the Zourdos study above), but the other cues are informal, so use them alongside your rep counting rather than instead of it. Our post on training to failure covers the trade-offs of going there regularly, so keep this routine to a handful of sets.

How hard should sets be?

The 2026 position stand from the American College of Sports Medicine (Currier et al., Medicine & Science in Sports & Exercise) pooled 137 systematic reviews covering more than 30,000 participants. Its primary recommendation is that healthy adults train with high effort at least twice a week, hitting all major muscle groups. It also states that taking sets to momentary failure does not enhance gains in strength, size or power and is not necessary. The authors suggest near-failure, or a target of about 2-3 reps in reserve, as a way to reach sufficient effort, while admitting there is not enough evidence to set exact RIR or RPE targets. We unpack the rest of the document in our guide to the ACSM guidelines.

Two meta-analyses help with the details. Refalo and colleagues (Sports Medicine, 2023) analyzed 15 studies on muscle size and found a trivial advantage for training to set failure over stopping short (effect size 0.19, confidence interval 0.00 to 0.37). For momentary failure specifically, the advantage was not distinguishable from zero (0.12, confidence interval -0.13 to 0.37). Robinson and colleagues (Sports Medicine, 2024) went further, estimating the reps in reserve in each study and running a series of meta-regressions. Across a wide range of RIR, the link to strength gains was negligible, but muscle growth improved as sets ended closer to failure. The authors describe their analysis as exploratory, partly because the RIR values had to be estimated from study descriptions.

Put together, the evidence says failure is optional, effort still matters, and the dial matters more for muscle size than for strength. Here is a practical translation. These ranges are suggestions built on the evidence above, not prescriptions tested in trials:

Heavy strength sets (1-6 reps) on squat, bench and deadlift: RPE 7-9. Strength gains looked similar across a wide range of RIR, so there is little reason to grind.
Main muscle-building compounds: RPE 7-9, with the final set at 9 if technique holds up.
Machines and isolation work: RPE 8-10. The last set can reach RPE 10 when the setup is stable and a miss is harmless.
First months of lifting: RPE 6-8, with form and consistency ahead of effort.

Effort and set count interact, so read this alongside our post on training volume.

A worked example: compounds vs isolation

Take a bench press target of 3 sets of 6-8 reps. Set one might be 80 kg for 8 at RPE 7, set two 80 kg for 7 at RPE 8, and set three 80 kg for 6 at RPE 9. The weight stays fixed while effort rises set by set. Next week, if you reach 8, 8 and 7 at those same efforts, add the smallest jump available. Our plate calculator shows what to load per side. Details on the lift are on the barbell bench press page.

An isolation lift works differently. On leg extensions you might do 3 sets of 10-12 at RPE 8, 9 and 9-10. The machine supports you, technique costs are low, and a missed rep carries little risk. Because estimates drift above about 12 reps, keep the target range in the 8-12 zone if you want the RPE number to mean something.

Day-to-day, RPE also lets you adjust load. If your first set at a target RPE 8 felt like a 6, add weight. If it felt like a 9.5, take a little off. Keep the changes small. Autoregulating like this did at least as well as fixed percentages in a small trial (Helms et al., Frontiers in Physiology, 2018): 21 resistance-trained men trained for 8 weeks, either with loads prescribed as percentages of one-rep max or with self-selected loads targeting RIR-based RPE ranges. In that trial, load was nudged by 2% for every half point of RPE outside the target range, which is a reasonable starting point rather than a proven rule. The RPE group gained 17.1 kg on squat and 10.7 kg on bench against 13.9 kg and 9.6 kg for the percentage group, differences that were not statistically significant, with similar changes in muscle thickness. For how this fits into a longer plan, see our post on periodization.

Recording effort when your log has no RPE field

RPE only helps if you can compare it over time. The simplest method is to write it next to the set in whatever you use, in a notes app, on paper, or in a spreadsheet: "80 x 8 @7." If that feels like too much, use a fixed stopping rule instead, such as "every working set ends with two reps left," and then compare weight and reps from week to week. When the rule stays constant, the same weight moving for more reps is real progress. Without it, eight grinding reps and eight comfortable reps look identical in a log.

Easy Reps does not record RPE or RIR, so the number lives in your head or in a separate note. What the app keeps is every set's weight and reps in your workout history, which you can filter by exercise or muscle group, plus automatic PR and estimated-1RM tracking. Hold your stopping rule steady and let a few weeks of history show whether the same weight is moving for more reps.

Common mistakes and who this fits less well

Rating warm-ups. RPE is for working sets.
Rating how heavy the bar felt, not how many reps were left. A heavy first rep can still be 4 reps from failure.
Trusting a far-from-failure guess. A "five left" is usually less reliable than a "one left."
Using RPE 10 singles to find your max. If you want a number, read our post on testing your 1RM safely.
Counting other limits as failure. If grip, breathing or discomfort ended the set, note that, since it does not mean your muscles were out of reps.

New lifters should lean on clean technique and a steady rise in weight, since their effort estimates are the noisiest. Older adults deserve a separate caveat: in a 2025 study (Gómez-Redondo et al., Experimental Gerontology) of 25 community-dwelling adults averaging 68 years old, participants stopped chest press sets at what they called 2, 4 and 6 reps in reserve. Their calls at 2 and 4 were off by about 2.1 and 1.6 reps, while their call at 6 was close. Reported RPE barely differed across the three conditions (8.0, 7.6 and 7.3). The authors concluded that predicted RIR may lack precision for prescribing exercise in older adults but might be useful for monitoring volume. With misses that large, older lifters should not lean on the "add a rep" rule of thumb above. The ACSM authors also note that lifting to failure may be inadvisable for some groups, such as older adults, because of vascular risk and form breakdown. If you have a heart condition, high blood pressure, or are returning from an injury, ask your clinician or physiotherapist how close to failure is appropriate for you. Regardless of age, stop and get medical advice if training ever brings on chest pain, fainting or near-fainting, or unusual breathlessness.

Bottom line

RPE and RIR are two names for one idea: how many reps you had left. Use the conversion (RPE 10 = 0 RIR, 9 = 1, 8 = 2, 7 = 3), expect to underestimate by about a rep on average, and trust your guess most when you are within two reps of the end of a set. Most working sets should finish with 1-3 reps in reserve, with machine and isolation sets allowed to go closer. Calibrate every few weeks on a safe lift by calling your "two left" and counting what is actually there. Then write the effort next to the set, or use a fixed stopping rule, so your log tells you what is really progressing.