Direct answer
Picture the last rep of a set of eight where the bar still moves like the first rep, and the lifter writes “failure” because the set felt serious. The notebook now says something that did not happen. Reps in reserve (RIR) is the plain question: how many more clean reps could you have done? Rate of perceived exertion (RPE), in the form lifters use, is that same question turned into a number.
In 2016, Zourdos and colleagues described a resistance-training RPE scale in the Journal of Strength and Conditioning Research where an RPE of 10 means no reps left, an RPE of 9 means about one rep left, an RPE of 8 means about two, and so on. They studied 29 people squatting, some experienced and some novice, and called the scale a practical way to regulate daily training load. The American College of Sports Medicine’s 2026 position stand says taking sets to momentary failure does not enhance strength, hypertrophy, or power, so you do not have to live at RPE 10. The same stand says there is not enough evidence to name one exact RIR or RPE number as the only correct target. Use the scale as a description. Do not use it as a moral score.
The scale, without the folklore
| RPE on the Zourdos-style scale | Reps you think you still had | What the set probably looked like |
|---|---|---|
| 10 | 0 | Another clean rep was not there. This is a miss or the rep just before one. |
| 9 | about 1 | The last rep was hard and slower, and one more might have existed. |
| 8 | about 2 | Hard, still similar in shape to the earlier reps. |
| 7 | about 3 | You stopped while the set still looked tidy. |
| 6 and below | several | A warm-up or an easy day, unless the load was supposed to be easy. |
ACSM’s authors propose that sufficient effort can be reached near failure or around a target of about 2–3 repetitions in reserve. On this scale, about 2–3 RIR lands near RPE 7 to 8. That is a useful default for a lot of work sets. It is not a law. Some sets in a strength block will sit closer to RPE 8–9. A deload should sit easier than that. The error is writing RPE 9 on a set you could have continued for four reps, because you want the log to look fierce.
Zourdos and colleagues also found that experienced squatters recorded a higher RPE at a one-rep max than the novice squatters did. If you are new, your number will be rougher. Treat it as a note you will get better at, and film a set now and then so the note has a picture. Do not wait for a perfect rating before you train.
How to assign the number after the set
Finish the set. Rack the bar. Then ask the question. Asking during the rep makes people either stop too early or grind a rep they cannot grade. The grade is about clean reps, not about a rep that needed a new back shape, a bounce, or a spotter doing half the work. If the last rep was a different exercise, the RPE is lying. Write the reps that stayed honest, and write a flag. Tracking workouts without obsession is the card those fields belong on.
Compare the number with bar speed if you have nothing but your eyes. Zourdos’s squatters showed an inverse relationship between average bar velocity and the RPE they reported: as the bar got slower in the ways the study measured, RPE went up. You do not need a velocity device to use that idea. If the last rep crawled and you wrote “RPE 6,” believe the crawl.
You might be thinking you should add load every time a set is “only” RPE 7. Sometimes yes. Sometimes the next step is another rep at the same load, or simply repeating the set because last week’s depth was sloppy and this week’s was not. Progressive overload without ego lifting is the decision. RPE tells you whether there was room. It does not tell you that room must be spent today.
Where the number helps, and where it gets in the way
It helps when two weeks look identical on paper — same load, same reps — and the effort note shows the second week was a grind. That is a reason to hold the load or to take an easier week, not a reason to “beat the log.” How to run a deload week that actually helps uses that kind of slide as one sign to pull back.
It helps on accessories, where a percentage of your squat max is a fantasy. You can still write that a row was about 2 reps from a miss. ACSM points out that methods based on proximity to failure are appealing precisely because they work for bands, machines, and bodyweight, not only for a barbell with a known max.
It gets in the way when you chase RPE 10 because a caption said growth only happens there. The ACSM overview does not support that. Failure is allowed on a set you choose. It is not required, and for some people — the stand mentions older adults as one example — grinding to a miss is a poor idea because form breaks. Natural lifters who train alone should be especially unwilling to miss a squat rep they cannot dump safely.
It gets in the way when the number replaces the video. A tidy RPE on a shorter range of motion is a tidy fiction. Full range of motion was one of the strength findings in the ACSM stand. Rate the set you meant to train.
A week that uses the scale without worshipping it
This is an illustrative pattern for one lift. It is not a result from the Zourdos study, which tested squats across set intensities rather than prescribing a month.
| Session | Work | Effort you are willing to see |
|---|---|---|
| Day A | 3 sets of 5 at a load you already own | Most sets around 2 reps in reserve. Stop if the third set becomes a different lift. |
| Day B | 2–3 sets of 8 | Leave about 2–3 reps. Add a rep next time only if all sets still match. |
| Easy week | Fewer sets, same exercise | Several reps left. The number should look easier than Day A on purpose. |
The one-rep max calculator can turn a hard set into a rough estimate when you want one. It is not more true than the RPE you felt. A calculator does not know that the last rep was a half rep. Your note does, if you told the truth.
Limits of a feeling
The scale is a rating you make. It drifts when you are new, when you are talking yourself up, and when the gym is loud. It is still better than no note. It will not tell you if a joint is injured. Pain is not an RPE. A sharp catch in a shoulder is a stop, not “RPE 9.” Get a qualified clinician involved when pain persists, when you already have an injury, or when a set comes with chest pain, dizziness, or numbness.
On the next set you care about, how many clean reps will you leave unused, and what will you write if the last rep stops looking like the first?
Sources
- Zourdos MC, Klemp A, Dolan C, et al. Novel resistance training–specific rating of perceived exertion scale measuring repetitions in reserve. Journal of Strength and Conditioning Research. 2016;30(1):267–275. Abstract record.
- Currier BS, D’Souza AC, Fiatarone Singh MA, et al. ACSM position stand on resistance-training prescription for healthy adults. Medicine & Science in Sports & Exercise. 2026. Failure and reps-in-reserve are discussed in the effort section. ACSM summary. Journal.
Educational material for healthy adults. An effort rating is not a medical test and not a personalized loading prescription.