A lifter grinds through a heavy barbell squat with visible effort while a training partner watches the bar

"Three sets of five at RPE 8" tells you more than "three sets of five at 100 kg" ever could, because it accounts for the day you slept badly and the day everything felt light. This guide covers what the numbers mean, which ones your sets should land on, how they translate to a percentage of your max, and how wrong your own ratings are likely to be.

What is RPE?

RPE stands for rate of perceived exertion: how hard a set felt, rated from 1 to 10. RIR stands for reps in reserve: how many more reps you could have done. They measure the same thing from opposite ends, and they add up to 10. Stop a set with two reps left and that is 2 RIR, which is RPE 8.

The lifting version of RPE is anchored to reps rather than to breathing. That matters, because you may have met a different RPE scale on a treadmill or in a class: the original Borg scale runs from 6 to 20 and was built to track heart rate during cardio. Same three letters, different tool. Everything below is the lifting scale.

The point of rating effort is that the right weight changes from day to day while the right effort does not. Your program says 100 kg (225 lb) for five, but after a bad night that set is a grind, and on a good day it flies. Prescribing the effort instead of the number lets the weight move to match the day. That idea has a name, autoregulation, and RPE is how you do it.

The RPE scale

RPEReps in reserveWhat it feels like
100Maximum effort. Nothing left, the bar stopped or would have
91Near maximum. One more rep was in there, and only one
82Very hard, but two clean reps were still available
73Hard. Clearly working, with a few reps to spare
64Somewhat hard. The last warm-up, or a light day
5 and below5 or moreEasy. Warm-ups and practice, not training

Half points matter more than they look, because most programs are written in them. RPE 8.5 means you were somewhere between one and two reps short and could not honestly say which. Use them whenever you are unsure rather than rounding to a number you do not mean.

One thing the table cannot tell you: RPE is about reps left in the tank, not about how unpleasant the set was. A set of 20 at RPE 7 will leave you breathing much harder than a heavy triple at RPE 9, and the triple is the closer of the two to failure. Rate the reps, not the discomfort.

RPE or RIR: which should you use?

Neither is better. They are the same measurement, so pick whichever your program speaks.

  • Programs written around heavy, low rep work usually use RPE, because it has half points and fine distinctions near the top matter when a single rep is a big jump in load.
  • Programs written around muscle growth usually use RIR, because "stop two reps short" is a plainer instruction when you are doing sets of 12.

If you are writing your own, use RIR. Counting reps you had left is a more concrete question than scoring effort out of 10, and concrete questions get more honest answers.

What RPE should you train at?

Nearly every working set belongs between RPE 7 and 9. Below 7 you are leaving too much on the table to drive much adaptation, and a steady diet of 10s buys fatigue faster than it buys progress.

Where you sit inside that band should depend on the exercise, and this is the part most guides skip:

Exercise typeTargetWhy
Heavy compounds (squat, deadlift, bench, overhead press)RPE 7 to 8Failure is risky under a loaded bar and costs days of recovery
Machine and supported compounds (leg press, chest press, hack squat)RPE 8 to 9Safe to push, still expensive in fatigue
Isolation (curls, lateral raises, leg extensions, calf work)RPE 8 to 10Failure is safe, cheap to recover from, and easier to judge

That last row is the one worth acting on. The fatigue cost of taking a lateral raise to failure is close to nothing, and your estimate of where failure is will be far better there than it is under a squat bar. Save your precision for the exercises where it is cheap.

By goal, the split is smaller than people expect. For strength, most work sits at RPE 7 to 8 with occasional heavier singles at 9. For muscle growth, RPE 8 to 9 across more total sets. Both goals live in the same band.

RPE as a percentage of your 1RM

Programs written in percentages and programs written in RPE are describing the same sets. You can convert between them, and almost nobody publishes a chart you can actually read, so here is one.

Find your rep count down the left, read across to the effort you are aiming for, and that is roughly the percentage of your one-rep max to load:

RepsRPE 10RPE 9RPE 8RPE 7RPE 6
1100%94%91%88%86%
294%91%88%86%83%
391%88%86%83%81%
488%86%83%81%79%
586%83%81%79%77%
683%81%79%77%75%
781%79%77%75%73%
879%77%75%73%71%
977%75%73%71%70%
1075%73%71%70%68%
1173%71%70%68%67%
1271%70%68%67%65%

So five reps at RPE 8 is about 81% of your max. If your bench max is 100 kg (225 lb), that is 81 kg (180 lb) for a set of five that leaves two in the tank.

These come from the same arithmetic behind an estimated one-rep max: a set of five with two reps left is a set of seven to failure, and seven reps to failure lands near 81% of max. Treat the numbers as a starting point rather than a law. Your own leverages, how much a rep slows down near failure, and the exercise itself all shift them by a few percent, and the chart gets less reliable above 10 reps where small differences in endurance matter more than strength does.

How accurate are your ratings, really?

Honestly? Less accurate than you would like, in a direction worth knowing: people consistently think they have fewer reps left than they do.

A 2022 review pooling twelve studies and 414 lifters found that people underpredicted their reps to failure by about one rep on average, and that lifters varied enough between themselves that being two or three reps out is unremarkable. In plain terms, the set you called RPE 9 was often an RPE 8.

Three findings make this practical:

  • Accuracy improves the closer you get to failure. When trained men squatting sets of around 16 reps were asked to call their reps remaining, they were out by more than five reps when judging from 5 reps left, against about two reps when judging from 1 left. "Could I do one more?" is a far easier question than "could I do five more?".
  • It gets better as the set goes on. When 259 coaches watched video of sets and called the reps remaining, they were out by 4.8 reps a third of the way in, 2.0 reps at two thirds, and 1.2 reps by the end. Early reps tell you very little.
  • Experience matters less than you would guess. Training status made no meaningful difference in the review above, and neither did whether the lift was upper or lower body. Coaching experience barely moved the needle in the video study either, worth about a fifth of a rep per decade on the job.

That last point cuts against the common advice to avoid RPE until you have years of training behind you. The evidence does not really support waiting. It supports stopping closer to failure, and practising.

How to calibrate yourself

Ratings get honest through evidence, not time. Three things fix most of the drift:

  1. Take something to true failure, on purpose. Pick a safe isolation exercise, a cable pushdown or a leg extension, and keep going until the rep genuinely does not complete. That is RPE 10. Until you have felt it recently, every other number on the scale is a guess anchored to nothing.
  2. Predict, then test. On a set where you planned to stop at 2 RIR, keep going instead and count the reps you actually had. Were there two, or four? Do this once every couple of weeks on a safe exercise and your estimates sharpen quickly.
  3. Film a set. Rep speed is the honest witness. If your last rep moved as fast as your first, you were further from failure than it felt, whatever the number you wrote down.

The most common failure mode is the opposite of ego: people rate sets as harder than they were, stop short, and wonder why nothing is moving. If every set in your log says RPE 9 and your lifts have not changed in two months, the ratings are the thing to question first.

Rate a set in two taps

PRZone puts an effort rating on any set, shows you what you rated it last time, and charts how your effort is trending for each exercise. You can rate it from your Apple Watch too. Free to start, no ads.

Reading your effort back

Logging RPE is only worth the taps if you look at it later. Almost every guide tells you to record it and then stops there, which is backwards: a single rating is noise, and the pattern across weeks is the whole value.

Two readings are worth your attention:

Effort rising while the weight stays flat. The same 100 kg for the same five reps felt like an 8 last month and feels like a 9 now. Nothing on your log changed, but you are heading toward a stall. This is the earliest warning you get, and it shows up well before your reps actually drop. It usually means fatigue is accumulating, which makes it a deload signal rather than a reason to push harder.

Effort falling while the weight stays flat. The same work feels easier than it did. That is adaptation, and it is your cue to add weight. It is also the only way to see progress during a stretch where the numbers on the bar have not moved, which is most stretches.

In PRZone the effort rating lives on the set row: tap the column header to switch it from Previous to Effort, and the picker shows a short description for every number so you are not guessing what a 7 means. The Effort trend for each exercise averages your working sets, leaves warm-ups out so their easy ratings cannot flatten the line, and counts sets taken to failure at full value.

Common mistakes

  • Rating the burn instead of the reps. High rep sets feel awful long before they are near failure. The question is always how many more you could have done.
  • Writing 9 on everything. If every set gets the same number, the scale has stopped carrying information and you may as well not log it.
  • Treating RPE as a target to beat. It describes a set, it is not a score. A session of honest 7s during a hard week is better training than a session of dishonest 9s.
  • Chasing decimals. The difference between 2 and 2.5 reps in reserve is well inside your own margin of error. Get the whole number roughly right and move on.
  • Using it on your first week of a new exercise. You cannot judge how close to failure you are on a movement you cannot yet do well. Learn the pattern first, rate it after.
  • Never going to failure at all. If you have not felt RPE 10 in six months, your 9s have quietly drifted down to 7s.

Frequently asked questions

What does RPE mean in the gym?

RPE stands for rate of perceived exertion: how hard a set felt, rated from 1 to 10. In lifting the top of the scale is anchored to reps rather than to breathing, so RPE 10 means you could not have done another rep and RPE 8 means you had about two left. The cardio version of RPE you may have seen, which runs from 6 to 20 and tracks heart rate, is a different scale for a different job.

What is the difference between RPE and RIR?

They are the same measurement counted from opposite ends. RIR counts the reps you had left, RPE counts how close you got to failure, and the two add up to 10. Three sets of 5 at RPE 8 and three sets of 5 at 2 RIR are identical instructions. Use whichever your program is written in.

What RPE should I train at?

Most working sets belong at RPE 7 to 9. Compound lifts like squats, deadlifts and presses are best kept at RPE 7 to 8, roughly 2 to 3 reps in reserve, because failing them is risky and expensive to recover from. Isolation work like curls, lateral raises and leg extensions can safely run at RPE 8 to 10. Anything below RPE 6 is a warm-up.

What does RPE 8 mean?

RPE 8 means you stopped with about 2 reps in reserve: hard and clearly effortful, but you could have done two more clean reps if you had to. It is the most commonly prescribed effort in strength programs because it is close enough to failure to drive adaptation and far enough away to repeat several times in a session.

Should beginners use RPE?

Yes, but expect your early ratings to be wrong, and do not build your whole program around them in the first few months. The useful beginner version is simpler than the scale: finish each set knowing you could have done one or two more clean reps. Rate it anyway, because practice at judging sets is the only thing that makes the ratings accurate.

How long does it take to learn RPE?

Published advice ranges from a few weeks to not bothering for a couple of years, and none of it comes from a study that measured learning time. What the research does show is that accuracy depends more on where you stop than on how long you have trained: everyone judges a set better near failure than far from it, and training experience makes surprisingly little difference. Taking a few isolation sets to true failure teaches you the top of the scale faster than waiting does.

Track it in PRZone

Log every set, see last session's numbers while you lift, and catch PRs the moment they happen. Free to start on iPhone and Android, with no ads.

Training guide

Progressive Overload: How to Apply It and Know It Works

The seven ways to make training harder, which one to change first, the three numbers that prove it is working, and what to do when you stall.

12 min read
Workout routine

Push Pull Legs Routine: 3 and 6 Day Plans

A complete push pull legs routine with 3 and 6 day schedules, sets and reps for every exercise, and a simple way to add weight week after week.

8 min read
Workout routine

Upper Lower Split: 4 Day Workout Plan

A 4 day upper lower split with two upper and two lower workouts, sets and reps for every exercise, and a simple plan for adding weight.

5 min read